{"name":"AI Change Desk change signal feed","canonical":"https://michaelhbm.com/AiChangeDesk/changes","feedUrl":"https://michaelhbm.com/AiChangeDesk/discovery/changes.json","description":"The living source-linked signal ledger behind the AI Change Desk change tracker. Documented records point to episode files, episode syntheses also expose their cited public sources, research records point to sources under evaluation, and incomplete classifications remain explicitly visible.","sourceLastModified":"2026-08-10","ledgerLastModified":"2026-08-10","corpusBoundary":{"frozenThrough":"2026-04-29","livingLedgerThrough":"2026-08-10","pendingNormalization":0},"count":99,"counts":{"documented":55,"research":44,"episodeSyntheses":16,"withSourceLinks":99,"withoutSourceLinks":0,"recentEpisodeCoverage":0,"pendingSignalNormalization":0},"statusDefinitions":{"documented":"An operating implication already covered in the named published AI Change Desk episode file.","episodeSynthesis":"An episode-level operating synthesis dated by publication, linked to its native episode file and the public citations stored with that episode.","research":"A source-linked editorial lead still being evaluated; inclusion is not a claim of independent verification or endorsement.","pendingSignalNormalization":"A published episode that remains discoverable but is excluded from the living ledger until every required editorial field is present."},"themes":[{"id":"governance","label":"Governance","title":"Governance became an operating constraint","operatingQuestion":"Can we swap vendors, export evidence, and explain our risk tiers under pressure?"},{"id":"access","label":"Access","title":"Access became the primary agent risk","operatingQuestion":"Who can let an AI system act, and who can pause it immediately?"},{"id":"deployment","label":"Deployment","title":"Deployment choices now change audit burden","operatingQuestion":"Are we choosing for speed, or for controllable continuity?"},{"id":"security","label":"Security","title":"Security workflows need named ownership","operatingQuestion":"Who owns triage, patch approval, operator communication, and rollback in the same workflow?"},{"id":"validation","label":"Validation","title":"Validation moved closer to the release surface","operatingQuestion":"What has to pass before a model or agent change is allowed to scale?"},{"id":"ecosystem","label":"Ecosystem","title":"The ecosystem is thickening around control","operatingQuestion":"Are we buying software, or committing to an operating model?"}],"recentEpisodeCoverage":[],"signals":[{"id":"signal-documented-2026-08-10-ep040-when-a-prompt-becomes-a-file","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-08-10-ep040-when-a-prompt-becomes-a-file","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-08-10","tag":"EP040","group":"Published AI change management operating checks","title":"EP040: When a Prompt Becomes a File","analysis":"A familiar AI interface can change the control object without changing the user intent. Build a receipt for the resulting object, context, controls, lifecycle, and communication.","themes":["governance","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP040: When a Prompt Becomes a File","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep040-when-a-prompt-becomes-a-file"},{"type":"public-source","label":"OpenAI, ChatGPT Enterprise & Edu release notes","url":"https://help.openai.com/en/articles/10128477-chatgpt-enterprise-edu-release-notes"},{"type":"public-source","label":"OpenAI, ChatGPT Voice","url":"https://help.openai.com/en/articles/20001274"},{"type":"public-source","label":"OpenAI, ChatGPT Work and Codex","url":"https://help.openai.com/en/articles/20001275"},{"type":"public-source","label":"OpenAI, Managing workspace settings in ChatGPT Enterprise","url":"https://help.openai.com/en/articles/8411955"},{"type":"public-source","label":"European Data Protection Board, anonymisation and generative-AI web-scraping guidance announcement","url":"https://www.edpb.europa.eu/news/edpb-sheds-light-on-anonymisation-and-web-scraping-for-generative-ai-and-adopts-final-version_en"}]},{"id":"signal-documented-2026-08-03-ep039-whose-account-did-the-agent-use","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-08-03-ep039-whose-account-did-the-agent-use","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-08-03","tag":"EP039","group":"Published AI change management operating checks","title":"EP039: Whose Account Did the Agent Use?","analysis":"A privacy-forward delegated-identity check for proving whose credential and authority moved data through a connected AI workflow.","themes":["governance","access","deployment","ecosystem"],"sources":[{"type":"published-episode","label":"EP039: Whose Account Did the Agent Use?","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep039-whose-account-did-the-agent-use"},{"type":"public-source","label":"OpenAI, ChatGPT workspace agents for Enterprise and Business","url":"https://help.openai.com/en/articles/20001143-chatgpt-workspace-agents-for-enterprise-and-business"},{"type":"public-source","label":"Microsoft, Least privilege for AI agents","url":"https://learn.microsoft.com/en-us/security/zero-trust/sfi/least-privilege-for-ai-agents"},{"type":"public-source","label":"GitHub, Agentic audit log events","url":"https://docs.github.com/en/enterprise-cloud@latest/copilot/reference/agentic-audit-log-events"},{"type":"public-source","label":"European Commission, Guidelines on transparency obligations for providers and deployers of certain AI systems","url":"https://digital-strategy.ec.europa.eu/en/policies/guidelines-transparency-ai-generated-content"},{"type":"public-source","label":"European Commission, Transparency obligations under Article 50 of the AI Act","url":"https://digital-strategy.ec.europa.eu/en/faqs/transparency-obligations-under-article-50-ai-act"},{"type":"public-source","label":"European Commission, Navigating the AI Act","url":"https://digital-strategy.ec.europa.eu/en/faqs/navigating-ai-act"},{"type":"public-source","label":"OpenAI, Health in ChatGPT","url":"https://openai.com/index/health-in-chatgpt/"},{"type":"public-source","label":"OpenAI, Health in ChatGPT help documentation","url":"https://help.openai.com/en/articles/20001036-health-in-chatgpt"},{"type":"public-source","label":"OpenAI, Health Privacy Policy","url":"https://openai.com/policies/health-privacy-policy/"},{"type":"public-source","label":"European Data Protection Board, Anonymisation and web scraping for generative AI","url":"https://www.edpb.europa.eu/news/edpb-sheds-light-on-anonymisation-and-web-scraping-for-generative-ai-and-adopts-final-version_en"}]},{"id":"signal-documented-2026-07-28-ep038-the-receipt-is-the-trajectory","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-07-28-ep038-the-receipt-is-the-trajectory","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-07-28","tag":"EP038","group":"Published AI change management operating checks","title":"EP038: The Receipt Is the Trajectory","analysis":"OpenAI and Hugging Face evaluation signals become a practical trajectory-receipt check: what changed, what evidence remains, and who owns the next decision.","themes":["governance","security","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP038: The Receipt Is the Trajectory","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep038-the-receipt-is-the-trajectory"},{"type":"public-source","label":"OpenAI, OpenAI and Hugging Face partner to address security incident during model evaluation","url":"https://openai.com/index/hugging-face-model-evaluation-security-incident/"},{"type":"public-source","label":"Hugging Face, Security incident disclosure - July 2026","url":"https://huggingface.co/blog/security-incident-july-2026"},{"type":"public-source","label":"OpenAI, Safety and alignment in an era of long-horizon models","url":"https://openai.com/index/safety-alignment-long-horizon-models/"},{"type":"public-source","label":"OpenAI, Introducing OpenAI Presence","url":"https://openai.com/index/introducing-openai-presence/"},{"type":"public-source","label":"OpenAI, Launching Health in ChatGPT","url":"https://openai.com/index/health-in-chatgpt/"},{"type":"public-source","label":"Associated Press incident reporting","url":"https://apnews.com/article/openai-gpt56-sol-hugging-face-63ab84fed5612af04d8a160d60f6def3"}]},{"id":"signal-documented-2026-07-20-ep037-work-agent-receipt-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-07-20-ep037-work-agent-receipt-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-07-20","tag":"EP037","group":"Published AI change management operating checks","title":"EP037: Work Agent Receipt Check","analysis":"When an AI agent says the work is finished, what receipt proves the right outcome was delivered at an acceptable total cost, under approved access, evidence, and review conditions?","themes":["governance","access","deployment","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP037: Work Agent Receipt Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep037-work-agent-receipt-check"},{"type":"public-source","label":"OpenAI, A scorecard for the AI age","url":"https://openai.com/index/a-scorecard-for-the-ai-age/"},{"type":"public-source","label":"OpenAI Help Center, ChatGPT - Release Notes","url":"https://help.openai.com/en/articles/6825453-chatgpt-release-notes"},{"type":"public-source","label":"OpenAI Help Center, ChatGPT Work and Codex","url":"https://help.openai.com/en/articles/20001275/"},{"type":"public-source","label":"Microsoft Learn, Transition Microsoft Copilot Studio and Microsoft Foundry agent security capabilities to Microsoft Agent 365","url":"https://learn.microsoft.com/en-us/defender-xdr/security-for-ai/transition-agent-security-to-agent-365"},{"type":"public-source","label":"Microsoft Learn, What's new in Copilot Studio","url":"https://learn.microsoft.com/en-us/microsoft-copilot-studio/whats-new"}]},{"id":"signal-documented-2026-06-29-ep036-preview-before-power-mode","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-06-29-ep036-preview-before-power-mode","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-06-29","tag":"EP036","group":"Published AI change management operating checks","title":"EP036: Preview Before Power Mode","analysis":"OpenAI's GPT-5.6 Sol preview shifts the operating question from model hype to frontier access control: who gets the strongest capability, where it runs, what it touches, what evidence remains, and who can roll it back before preview power becomes normal work.","themes":["governance","access","deployment","security","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP036: Preview Before Power Mode","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep036-preview-before-power-mode"},{"type":"public-source","label":"OpenAI","url":"https://openai.com/index/previewing-gpt-5-6-sol/"},{"type":"public-source","label":"OpenAI Help Center","url":"https://help.openai.com/en/articles/20001325-a-preview-of-gpt-56-sol-terra-and-luna"},{"type":"public-source","label":"OpenAI agents/work research","url":"https://openai.com/index/how-agents-are-transforming-work/"},{"type":"public-source","label":"Microsoft Copilot in Excel","url":"https://www.microsoft.com/en-us/microsoft-365/blog/2026/06/25/copilot-in-excel-built-for-the-era-of-frontier-finance/"},{"type":"public-source","label":"Anthropic Claude Tag","url":"https://www.anthropic.com/news/introducing-claude-tag"}]},{"id":"signal-documented-2026-06-22-ep034-patch-before-prod","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-06-22-ep034-patch-before-prod","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-06-22","tag":"EP034","group":"Published AI change management operating checks","title":"EP034: Patch Before Prod","analysis":"OpenAI Daybreak and Patch the Planet move AI security work from finding bugs toward patch-chain ownership: validation, approval, tests, rollback, disclosure, budget, and replacement before fixes touch production.","themes":["governance","deployment","security","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP034: Patch Before Prod","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep034-patch-before-prod"},{"type":"public-source","label":"OpenAI Daybreak","url":"https://openai.com/index/daybreak-securing-the-world/"},{"type":"public-source","label":"OpenAI Patch the Planet","url":"https://openai.com/index/patch-the-planet/"},{"type":"public-source","label":"OpenAI enterprise spend controls","url":"https://openai.com/index/chatgpt-enterprise-spend-controls/"},{"type":"public-source","label":"OpenAI ChatGPT release notes","url":"https://help.openai.com/en/articles/6825453-chatgpt-release-notes"},{"type":"public-source","label":"Microsoft Copilot Cowork GA","url":"https://www.microsoft.com/en-us/microsoft-365/blog/2026/06/16/copilot-cowork-is-now-generally-available/"},{"type":"public-source","label":"Microsoft Work IQ APIs","url":"https://www.microsoft.com/en-us/microsoft-365/blog/2026/06/02/announcing-the-new-work-iq-apis/"},{"type":"public-source","label":"Anthropic Fable/Mythos access statement","url":"https://www.anthropic.com/news/fable-mythos-access"},{"type":"public-source","label":"YouTube AI labels update","url":"https://blog.youtube/news-and-events/improving-ai-labels-viewers-creators/"},{"type":"public-source","label":"Podnews on RSS AI disclosure flag","url":"https://podnews.net/update/bumper-free"}]},{"id":"signal-documented-2026-06-17-ep033-agent-runtime-budget-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-06-17-ep033-agent-runtime-budget-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-06-17","tag":"EP033","group":"Published AI change management operating checks","title":"EP033: Agent Runtime Budget Check","analysis":"Microsoft Copilot Cowork and Work IQ make agent work a runtime budget question, while Anthropic access changes reinforce fallback planning before workflows depend on one model surface.","themes":["governance","access","deployment","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP033: Agent Runtime Budget Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep033-agent-runtime-budget-check"},{"type":"public-source","label":"Microsoft Copilot Cowork GA","url":"https://www.microsoft.com/en-us/microsoft-365/blog/2026/06/16/copilot-cowork-is-now-generally-available/"},{"type":"public-source","label":"Microsoft Work IQ APIs","url":"https://www.microsoft.com/en-us/microsoft-365/blog/2026/06/02/announcing-the-new-work-iq-apis/"},{"type":"public-source","label":"Anthropic Fable/Mythos access-removal statement","url":"https://www.anthropic.com/news/fable-mythos-access"},{"type":"public-source","label":"OpenAI ChatGPT release notes","url":"https://help.openai.com/en/articles/6825453-chatgpt-release-notes"},{"type":"public-source","label":"OpenAI Memory FAQ","url":"https://help.openai.com/en/articles/8590148-memory-faq/"}]},{"id":"signal-documented-2026-06-15-ep032-memory-summary-exit-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-06-15-ep032-memory-summary-exit-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-06-15","tag":"EP032","group":"Published AI change management operating checks","title":"EP032: Memory Summary Exit Check","analysis":"A practical check for proving what was corrected, deleted, suppressed, or still able to reappear when AI memory becomes a synthesized control surface.","themes":["governance","access","deployment","security","validation"],"sources":[{"type":"published-episode","label":"EP032: Memory Summary Exit Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep032-memory-summary-exit-check"},{"type":"public-source","label":"OpenAI ChatGPT release notes","url":"https://help.openai.com/en/articles/6825453-chatgpt-release-notes"},{"type":"public-source","label":"OpenAI Memory FAQ","url":"https://help.openai.com/en/articles/8590148-memory-faq/"},{"type":"public-source","label":"OpenAI Codex browser docs","url":"https://developers.openai.com/codex/app/browser"},{"type":"public-source","label":"OpenAI Lockdown Mode","url":"https://help.openai.com/en/articles/20001061-lockdown-mode"},{"type":"public-source","label":"OpenAI memory product post","url":"https://openai.com/index/chatgpt-memory-dreaming/"},{"type":"public-source","label":"YouTube AI labels update","url":"https://blog.youtube/news-and-events/improving-ai-labels-viewers-creators/"},{"type":"public-source","label":"Podnews AI disclosure guidance","url":"https://podnews.net/update/ai-disclosures"}]},{"id":"signal-documented-2026-06-08-ep031-memory-control-plane-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-06-08-ep031-memory-control-plane-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-06-08","tag":"EP031","group":"Published AI change management operating checks","title":"EP031: Memory Control Plane Check","analysis":"AI memory is becoming more useful, but usefulness creates a new operating surface. If the system can carry context forward, teams need a memory control plane: summary, source, correction, deletion, sensitive-work mode, and disclosure.","themes":["governance","access","deployment","security","validation"],"sources":[{"type":"published-episode","label":"EP031: Memory Control Plane Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep031-memory-control-plane-check"},{"type":"public-source","label":"OpenAI memory rollout","url":"https://openai.com/index/chatgpt-memory-dreaming/"},{"type":"public-source","label":"OpenAI Memory FAQ","url":"https://help.openai.com/en/articles/8590148-memory-faq/"},{"type":"public-source","label":"ChatGPT release notes","url":"https://help.openai.com/en/articles/6825453-chatgpt-release-notes"},{"type":"public-source","label":"Lockdown Mode","url":"https://help.openai.com/en/articles/20001061-lockdown-mode"},{"type":"public-source","label":"YouTube AI labels","url":"https://blog.youtube/news-and-events/improving-ai-labels-viewers-creators/"},{"type":"public-source","label":"Podnews AI disclosure guide","url":"https://podnews.net/update/ai-disclosures"}]},{"id":"signal-documented-2026-06-03-ep030-always-on-agent-control-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-06-03-ep030-always-on-agent-control-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-06-03","tag":"EP030","group":"Published AI change management operating checks","title":"EP030: Always-On Agent Control Check","analysis":"Microsoft and OpenAI moved agent work closer to persistent identity, internal apps, and publishable work surfaces. The operator question is who owns the standing permission before an agent keeps acting in the background.","themes":["governance","access","deployment","security","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP030: Always-On Agent Control Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep030-always-on-agent-control-check"},{"type":"public-source","label":"Episode show notes and source links","url":"https://rss.com/podcasts/aichangedesk/2882638"},{"type":"public-source","label":"AI Change Desk RSS feed","url":"https://media.rss.com/aichangedesk/feed.xml"},{"type":"public-source","label":"AI Change Desk episode archive","url":"https://michaelhbm.com/AiChangeDesk/episodes"}]},{"id":"signal-documented-2026-06-01-ep029-agent-reliability-evidence-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-06-01-ep029-agent-reliability-evidence-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-06-01","tag":"EP029","group":"Published AI change management operating checks","title":"EP029: Agent Reliability Evidence Check","analysis":"Agentic systems are getting longer leashes. This episode gives operators a five-receipt evidence check for deciding when agent workflows are reliable enough to scale.","themes":["governance","deployment","security","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP029: Agent Reliability Evidence Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep029-agent-reliability-evidence-check"},{"type":"public-source","label":"OpenAI ChatGPT release notes","url":"https://help.openai.com/en/articles/6825453-chatgpt-release-notes"},{"type":"public-source","label":"OpenAI Codex cloud documentation","url":"https://developers.openai.com/codex/cloud/"},{"type":"public-source","label":"Anthropic Claude Opus 4.8","url":"https://www.anthropic.com/news/claude-opus-4-8"},{"type":"public-source","label":"AWS LLM observability","url":"https://aws.amazon.com/blogs/machine-learning/comprehensive-observability-for-amazon-sagemaker-ai-llm-inference-from-gpu-utilization-to-llm-quality/"},{"type":"public-source","label":"AWS deep-agent evaluations","url":"https://aws.amazon.com/blogs/machine-learning/evaluating-deep-agents-using-langsmith-on-aws/"},{"type":"public-source","label":"AWS agent test-suite datasets","url":"https://aws.amazon.com/blogs/machine-learning/build-a-test-suite-that-grows-with-your-agent-with-dataset-management-in-amazon-bedrock-agentcore/"}]},{"id":"signal-documented-2026-05-25-ep027-no-new-delta-verification-discipline-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-05-25-ep027-no-new-delta-verification-discipline-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-05-25","tag":"EP027","group":"Published AI change management operating checks","title":"EP027: No-New-Delta Verification Discipline Check","analysis":"A Memorial Day release on no-new-delta verification discipline: how to report unchanged official evidence without inventing novelty, and why quiet source days still need a clean operating check.","themes":["governance","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP027: No-New-Delta Verification Discipline Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep027-no-new-delta-verification-discipline-check"},{"type":"public-source","label":"OpenAI ChatGPT release notes","url":"https://help.openai.com/en/articles/6825453-release-notes"},{"type":"public-source","label":"OpenAI provenance post","url":"https://openai.com/index/advancing-content-provenance/"},{"type":"public-source","label":"Podnews Report Card 2026 Results","url":"https://podnews.net/article/report-card-2026-results"},{"type":"public-source","label":"YouTube news from Google I/O 2026","url":"https://blog.youtube/news-and-events/youtube-news-google-io-2026/"},{"type":"public-source","label":"U.S. Census Bureau Memorial Day 2026","url":"https://www.census.gov/newsroom/stories/memorial-day.html"}]},{"id":"signal-documented-2026-05-20-ep026-agent-toolchain-ownership-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-05-20-ep026-agent-toolchain-ownership-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-05-20","tag":"EP026","group":"Published AI change management operating checks","title":"EP026: Agent Toolchain Ownership Check","analysis":"Agent announcements are becoming toolchain ownership problems: connectors, managed agents, SDKs, mobile approvals, evidence, fallback, and off switches.","themes":["governance","access","deployment","security","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP026: Agent Toolchain Ownership Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep026-agent-toolchain-ownership-check"},{"type":"public-source","label":"Anthropic: Anthropic acquires Stainless —","url":"https://www.anthropic.com/news/anthropic-acquires-stainless"},{"type":"public-source","label":"Anthropic: KPMG integrates Claude across its core business and workforce of more than 276,000 —","url":"https://www.anthropic.com/news/anthropic-kpmg"},{"type":"public-source","label":"Google: I/O 2026 collection —","url":"https://blog.google/innovation-and-ai/technology/developers-tools/google-io-2026-collection/"},{"type":"public-source","label":"Google: I/O 2026 developer highlights —","url":"https://blog.google/innovation-and-ai/technology/developers-tools/google-io-2026-developer-highlights/"},{"type":"public-source","label":"Google: I/O 2026 opening keynote —","url":"https://blog.google/innovation-and-ai/sundar-pichai-io-2026/"},{"type":"public-source","label":"OpenAI: Work with Codex from anywhere —","url":"https://openai.com/index/work-with-codex-from-anywhere/"},{"type":"public-source","label":"OpenAI: ChatGPT release notes —","url":"https://help.openai.com/en/articles/6825453-chatgpt-release-notes"}]},{"id":"signal-documented-2026-05-18-ep025-away-mode-control-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-05-18-ep025-away-mode-control-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-05-18","tag":"EP025","group":"Published AI change management operating checks","title":"EP025: Away-Mode Control Check","analysis":"AI work is following operators across phones, browsers, devices, sensitive connectors, and deployment programs. This episode turns that surface expansion into an away-mode control check: what can move while the owner is away, what must wait, what creates evidence, and who can stop it.","themes":["governance","access","deployment","security","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP025: Away-Mode Control Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep025-away-mode-control-check"},{"type":"public-source","label":"OpenAI, “Work with Codex from anywhere,” May 14, 2026.","url":"https://openai.com/index/work-with-codex-from-anywhere/"},{"type":"public-source","label":"OpenAI, “Simplify your personal finances with ChatGPT,” May 15, 2026.","url":"https://openai.com/index/personal-finance-chatgpt/"},{"type":"public-source","label":"Google, “A smarter, more proactive Android with Gemini Intelligence,” May 12, 2026.","url":"https://blog.google/products-and-platforms/platforms/android/gemini-intelligence/"},{"type":"public-source","label":"Anthropic, “Expanding our partnership with PwC,” May 14, 2026.","url":"https://www.anthropic.com/news/pwc-expanded-partnership"},{"type":"public-source","label":"OpenAI, “OpenAI launches The OpenAI Deployment Company,” May 11, 2026.","url":"https://openai.com/index/openai-launches-the-deployment-company/"}]},{"id":"signal-documented-2026-05-06-ep024-delegation-quality-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-05-06-ep024-delegation-quality-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-05-06","tag":"EP024","group":"Published AI change management operating checks","title":"EP024: Delegation Quality Check","analysis":"Agents are moving from answering questions to taking assignments, so the operating control is delegation quality: handoff, evidence, review, and fallback.","themes":["governance","access","deployment","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP024: Delegation Quality Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep024-delegation-quality-check"},{"type":"public-source","label":"Microsoft: Copilot Cowork: From conversation to action across skills, integrations, and devices -","url":"https://www.microsoft.com/en-us/microsoft-365/blog/2026/05/05/copilot-cowork-from-conversation-to-action-across-skills-integrations-and-devices/"},{"type":"public-source","label":"Microsoft: Microsoft Agent 365, now generally available, expands capabilities and integrations -","url":"https://www.microsoft.com/en-us/security/blog/2026/05/01/microsoft-agent-365-now-generally-available-expands-capabilities-and-integrations/"},{"type":"public-source","label":"Anthropic: Agents for financial services -","url":"https://www.anthropic.com/news/finance-agents"},{"type":"public-source","label":"Microsoft: Microsoft 365 Copilot, human agency, and the opportunity for every organization -","url":"https://www.microsoft.com/en-us/microsoft-365/blog/2026/05/05/microsoft-365-copilot-human-agency-and-the-opportunity-for-every-organization/"}]},{"id":"signal-documented-2026-05-04-ep023-trust-boundary-check","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-05-04-ep023-trust-boundary-check","lane":"documented","status":"Published AI Change Desk episode synthesis linked to its episode file and cited public sources","recordKind":"episode-synthesis","dateBasis":"episode-publication","date":"2026-05-04","tag":"EP023","group":"Published AI change management operating checks","title":"EP023: Trust Boundary Check","analysis":"Advanced Account Security, OpenAI on Amazon Bedrock, FedRAMP availability, partnership changes, and the May 8 macOS remediation deadline all point to one Monday operating question: when AI becomes infrastructure, who owns the trust boundary across identity, cloud channel, compliance scope, and endpoint evidence?","themes":["governance","access","deployment","security","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP023: Trust Boundary Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep023-trust-boundary-check"},{"type":"public-source","label":"OpenAI, Introducing Advanced Account Security","url":"https://openai.com/index/advanced-account-security/"},{"type":"public-source","label":"AWS, Amazon Bedrock now offers OpenAI models, Codex, and Managed Agents","url":"https://aws.amazon.com/about-aws/whats-new/2026/04/bedrock-openai-models-codex-managed-agents/"},{"type":"public-source","label":"Amazon, OpenAI Models on Amazon Bedrock","url":"https://www.aboutamazon.com/news/aws/bedrock-openai-models"},{"type":"public-source","label":"OpenAI, The next phase of the Microsoft OpenAI partnership","url":"https://openai.com/index/next-phase-of-microsoft-partnership/"},{"type":"public-source","label":"OpenAI, OpenAI available at FedRAMP Moderate","url":"https://openai.com/index/openai-available-at-fedramp-moderate/"},{"type":"public-source","label":"OpenAI, Our response to the Axios developer tool compromise","url":"https://openai.com/index/axios-developer-tool-compromise/"}]},{"id":"signal-research-2026-04-29-aws-said-amazon-bedrock-will-add-openai-models-codex-and-bedrock-managed-agents-support-in-limited-preview","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-29-aws-said-amazon-bedrock-will-add-openai-models-codex-and-bedrock-managed-agents-support-in-limited-preview","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-29","tag":null,"group":"Enterprise operating models thickened","title":"AWS said Amazon Bedrock will add OpenAI models, Codex, and Bedrock Managed Agents support in limited preview.","analysis":"OpenAI capability is moving into an existing enterprise governance and spend surface instead of requiring a separate buying path, so teams need to decide whether model choice, agent routing, approval controls, and budget ownership now belong with the cloud platform team, the AI platform team, or both.","themes":["deployment","validation","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.aboutamazon.com/news/aws/bedrock-openai-models"}]},{"id":"signal-research-2026-04-28-anthropic-launched-claude-for-creative-work-with-connectors-across-adobe-figma-canva-and-prisma","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-28-anthropic-launched-claude-for-creative-work-with-connectors-across-adobe-figma-canva-and-prisma","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-28","tag":null,"group":"Control infrastructure kept moving","title":"Anthropic launched Claude for Creative Work with connectors across Adobe, Figma, Canva, and Prisma.","analysis":"Creative-production AI is becoming connector-native instead of file-export-first, so teams need app-scope review, artifact lineage, permission boundaries, rollback rules, and brand-approval checkpoints before generated or modified assets flow directly through design systems.","themes":["deployment","validation","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.anthropic.com/news/claude-for-creative-work"}]},{"id":"signal-documented-2026-04-27-openai-and-microsoft-announced-the-next-phase-of-their-partnership","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-04-27-openai-and-microsoft-announced-the-next-phase-of-their-partnership","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-27","tag":"EP022","group":"Access lifecycle became a release-control problem","title":"OpenAI and Microsoft announced the next phase of their partnership.","analysis":"Cloud and licensing assumptions are changing. The operating question is not vendor gossip; it is whether procurement, fallback, and dependency maps stay current when the underlying route to capability becomes less simple.","themes":["ecosystem","deployment","governance"],"sources":[{"type":"published-episode","label":"EP022: Access Lifecycle Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep022-access-lifecycle-check"}]},{"id":"signal-documented-2026-04-27-openai-announced-fedramp-moderate-availability-for-chatgpt-enterprise-and-the-api-platform","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-04-27-openai-announced-fedramp-moderate-availability-for-chatgpt-enterprise-and-the-api-platform","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-27","tag":"EP022","group":"Access lifecycle became a release-control problem","title":"OpenAI announced FedRAMP Moderate availability for ChatGPT Enterprise and the API Platform.","analysis":"Regulated access expanded, but approved access is not the same thing as approved use. Agencies and adjacent enterprises still need allowed-use mapping, evidence reuse rules, shared-responsibility clarity, and plain-language deployment boundaries.","themes":["governance","access","deployment","validation"],"sources":[{"type":"published-episode","label":"EP022: Access Lifecycle Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep022-access-lifecycle-check"}]},{"id":"signal-documented-2026-04-26-openai-s-sora-web-and-app-experience-reached-discontinuation-while-the-api-sunset-remains-set-for-september-24","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-04-26-openai-s-sora-web-and-app-experience-reached-discontinuation-while-the-api-sunset-remains-set-for-september-24","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-26","tag":"EP022","group":"Access lifecycle became a release-control problem","title":"OpenAI's Sora web and app experience reached discontinuation while the API sunset remains set for September 24.","analysis":"A product sunset is a workflow-control event. Teams need export completion checks, migration owners, archive decisions, deletion handling, and customer-delivery plans before the surface disappears.","themes":["access","deployment","validation"],"sources":[{"type":"published-episode","label":"EP022: Access Lifecycle Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep022-access-lifecycle-check"}]},{"id":"signal-research-2026-04-23-openai-launched-gpt-5-5-as-a-new-flagship-model-for-chatgpt-and-codex","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-23-openai-launched-gpt-5-5-as-a-new-flagship-model-for-chatgpt-and-codex","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-23","tag":null,"group":"Control infrastructure kept moving","title":"OpenAI launched GPT-5.5 as a new flagship model for ChatGPT and Codex.","analysis":"Frontier model upgrades are now arriving as day-one workflow changes across coding, research, computer use, and document-heavy work, so teams need prompt-harness regression checks, approval boundaries, token and spend monitoring, and rollback criteria before stronger autonomy becomes the default operating surface.","themes":["deployment","validation","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/introducing-gpt-5-5/"}]},{"id":"signal-documented-2026-04-22-openai-s-enterprise-and-edu-release-notes-added-workspace-agents-rollout-details-for-business-and-enterprise-workspaces","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-04-22-openai-s-enterprise-and-edu-release-notes-added-workspace-agents-rollout-details-for-business-and-enterprise-workspaces","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-22","tag":"EP022","group":"Access lifecycle became a release-control problem","title":"OpenAI's Enterprise and Edu release notes added Workspace Agents rollout details for Business and Enterprise workspaces.","analysis":"Scheduled, app-connected agents move AI usage from ad hoc prompting into repeatable operations. Teams need publication controls, app-scope review, analytics ownership, scheduled-run approval, and a clear off switch before agent workflows become normal work.","themes":["access","deployment","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP022: Access Lifecycle Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep022-access-lifecycle-check"}]},{"id":"signal-documented-2026-04-21-openai-announced-chatgpt-images-2-0-inside-the-everyday-chatgpt-work-surface","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-04-21-openai-announced-chatgpt-images-2-0-inside-the-everyday-chatgpt-work-surface","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-21","tag":"EP020","group":"Visual work became a production surface","title":"OpenAI announced ChatGPT Images 2.0 inside the everyday ChatGPT work surface.","analysis":"More capable image creation inside a general-purpose assistant means visual artifacts can move from prompt to public asset faster than review habits expect. Teams need source records, prompt records, copy checks, alt text, allowed-use notes, and rollback files before polished images ship.","themes":["deployment","validation","governance"],"sources":[{"type":"published-episode","label":"EP020: Visual Workflow Control Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep020-visual-workflow-control-check"}]},{"id":"signal-documented-2026-04-20-anthropic-and-amazon-expanded-their-collaboration-around-up-to-5-gigawatts-of-new-compute-capacity","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-04-20-anthropic-and-amazon-expanded-their-collaboration-around-up-to-5-gigawatts-of-new-compute-capacity","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-20","tag":"EP022","group":"Access lifecycle became a release-control problem","title":"Anthropic and Amazon expanded their collaboration around up to 5 gigawatts of new compute capacity.","analysis":"Decade-scale compute commitments make infrastructure capacity part of the operating model. Procurement and continuity owners need to know which workflows depend on which provider capacity and what fallback means if access, terms, or availability changes.","themes":["deployment","ecosystem"],"sources":[{"type":"published-episode","label":"EP022: Access Lifecycle Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep022-access-lifecycle-check"}]},{"id":"signal-documented-2026-04-17-anthropic-launched-claude-design-as-a-research-preview-visual-work-and-handoff-surface","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-04-17-anthropic-launched-claude-design-as-a-research-preview-visual-work-and-handoff-surface","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-17","tag":"EP020","group":"Visual work became a production surface","title":"Anthropic launched Claude Design as a research-preview visual work and handoff surface.","analysis":"Designs, decks, prototypes, exports, organization sharing, and implementation handoffs now sit closer to AI-assisted production. That makes brand-system permission, export control, accessibility review, and handoff QA part of the visual workflow rather than a late polish step.","themes":["deployment","validation","ecosystem"],"sources":[{"type":"published-episode","label":"EP020: Visual Workflow Control Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep020-visual-workflow-control-check"}]},{"id":"signal-research-2026-04-16-anthropic-released-claude-opus-4-7-with-stronger-long-running-software-work-cyber-use-safeguards-xhigh-effort-control-and-task-budgets","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-16-anthropic-released-claude-opus-4-7-with-stronger-long-running-software-work-cyber-use-safeguards-xhigh-effort-control-and-task-budgets","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-16","tag":null,"group":"Control infrastructure kept moving","title":"Anthropic released Claude Opus 4.7 with stronger long-running software work, cyber-use safeguards, xhigh effort control, and task budgets.","analysis":"Frontier model upgrades are now shipping as operating changes, not just capability bumps: teams need prompt-harness retuning, token-budget checks, cyber-use approval boundaries, and release gates before harder autonomous coding or review work moves to the new model.","themes":["validation","security","deployment"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.anthropic.com/news/claude-opus-4-7"}]},{"id":"signal-research-2026-04-16-openai-named-initial-trusted-access-for-cyber-participants-and-a-10-million-cyber-defense-grant-path","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-16-openai-named-initial-trusted-access-for-cyber-participants-and-a-10-million-cyber-defense-grant-path","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-16","tag":null,"group":"Control infrastructure kept moving","title":"OpenAI named initial Trusted Access for Cyber participants and a $10 million cyber-defense grant path.","analysis":"High-capability defensive AI is moving into a trust-gated operating program, so security leaders need eligibility criteria, use-case approval boundaries, evidence-sharing rules, and escalation paths before advanced cyber access becomes a normal tool in vulnerability research and incident response.","themes":["security","validation","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/accelerating-cyber-defense-ecosystem/"}]},{"id":"signal-research-2026-04-15-jane-street-committed-about-6-billion-to-coreweave-s-ai-cloud-and-made-a-1-billion-equity-investment","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-15-jane-street-committed-about-6-billion-to-coreweave-s-ai-cloud-and-made-a-1-billion-equity-investment","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-15","tag":null,"group":"Control infrastructure kept moving","title":"Jane Street committed about $6 billion to CoreWeave's AI cloud and made a $1 billion equity investment.","analysis":"AI capacity pressure is spreading from frontier labs into quantitative finance, so infrastructure planning now needs customer-concentration checks, compute-allocation assumptions, and continuity scenarios for non-lab buyers competing for next-generation AI cloud capacity.","themes":["deployment","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.coreweave.com/news/jane-street-signs-6-billion-ai-cloud-agreement-with-coreweave"}]},{"id":"signal-research-2026-04-10-openai-disclosed-its-response-to-the-axios-developer-tool-compromise-affecting-chatgpt-and-codex-macos-app-signing-workflows","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-10-openai-disclosed-its-response-to-the-axios-developer-tool-compromise-affecting-chatgpt-and-codex-macos-app-signing-workflows","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-10","tag":null,"group":"Control infrastructure kept moving","title":"OpenAI disclosed its response to the Axios developer-tool compromise affecting ChatGPT and Codex macOS app signing workflows.","analysis":"Desktop AI access is now a software-supply-chain continuity issue, not just an endpoint patch task, so teams need upgrade deadlines, trusted-version inventories, and fallback access plans before expired signed builds quietly break local coding and review workflows.","themes":["security","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/axios-developer-tool-compromise/"}]},{"id":"signal-research-2026-04-08-nist-opened-its-concept-note-for-a-trustworthy-ai-in-critical-infrastructure-profile","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-08-nist-opened-its-concept-note-for-a-trustworthy-ai-in-critical-infrastructure-profile","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-08","tag":null,"group":"Policy and geographic footprint kept expanding","title":"NIST opened its concept note for a Trustworthy AI in Critical Infrastructure profile.","analysis":"Critical-infrastructure AI is moving toward a named operating profile instead of generic AI-governance language, so operators need to map trustworthiness requirements, supply-chain roles, and evidence expectations before high-stakes deployments harden around ad hoc controls.","themes":["governance","deployment","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.nist.gov/programs-projects/concept-note-ai-rmf-profile-trustworthy-ai-critical-infrastructure"}]},{"id":"signal-research-2026-04-08-openai-framed-the-next-phase-of-enterprise-ai-around-company-wide-agents-and-a-unified-work-surface","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-08-openai-framed-the-next-phase-of-enterprise-ai-around-company-wide-agents-and-a-unified-work-surface","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-08","tag":null,"group":"Enterprise operating models thickened","title":"OpenAI framed the next phase of enterprise AI around company-wide agents and a unified work surface.","analysis":"Enterprise rollout is being positioned as an operating-model transition instead of a loose tool pilot, so teams need stage-gate ownership, agent permissions, workflow logging, and budget guardrails before cross-system AI adoption gets ahead of governance.","themes":["deployment","validation","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/next-phase-of-enterprise-ai/"}]},{"id":"signal-research-2026-04-07-anthropic-launched-project-glasswing-with-mythos-preview-for-defensive-cybersecurity-operations","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-07-anthropic-launched-project-glasswing-with-mythos-preview-for-defensive-cybersecurity-operations","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-07","tag":null,"group":"Control infrastructure kept moving","title":"Anthropic launched Project Glasswing with Mythos Preview for defensive cybersecurity operations.","analysis":"Frontier cyber capability is moving into organized defensive workflows with named partners, credits, and funding, so security teams now need explicit access limits, escalation paths, and evidence rules before model-assisted vulnerability discovery outruns response ownership.","themes":["security","validation","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.anthropic.com/glasswing"}]},{"id":"signal-research-2026-04-06-anthropic-expanded-its-google-and-broadcom-partnership-around-multi-gigawatt-tpu-compute-capacity","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-06-anthropic-expanded-its-google-and-broadcom-partnership-around-multi-gigawatt-tpu-compute-capacity","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-06","tag":null,"group":"Control infrastructure kept moving","title":"Anthropic expanded its Google and Broadcom partnership around multi-gigawatt TPU compute capacity.","analysis":"Infrastructure concentration and capacity access are becoming frontline operating constraints, so deployment owners need scenario plans for vendor dependency, long-range cost exposure, and what happens if compute leverage shifts faster than control design does.","themes":["deployment","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.anthropic.com/news/google-broadcom-partnership-compute"}]},{"id":"signal-research-2026-04-05-openai-s-current-file-library-guidance-says-uploaded-chatgpt-files-remain-reusable-until-users-delete-them","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-05-openai-s-current-file-library-guidance-says-uploaded-chatgpt-files-remain-reusable-until-users-delete-them","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-05","tag":null,"group":"Vendor governance hardened at the legal and privacy layer","title":"OpenAI's current File Library guidance says uploaded ChatGPT files remain reusable until users delete them.","analysis":"File handling in ChatGPT is now a persistent workflow surface instead of a transient chat action, so teams need retention tiers, pre-share review steps, and explicit deletion guidance before reused files start carrying unmanaged policy or customer data forward.","themes":["deployment","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://help.openai.com/en/articles/20001052-file-storage-and-library-in-chatgpt"}]},{"id":"signal-research-2026-04-03-openai-documented-legacy-model-access-rules-for-enterprise-and-edu-users-after-the-gpt-4o-cutoff","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-03-openai-documented-legacy-model-access-rules-for-enterprise-and-edu-users-after-the-gpt-4o-cutoff","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-03","tag":null,"group":"Vendor governance hardened at the legal and privacy layer","title":"OpenAI documented legacy-model access rules for Enterprise and Edu users after the GPT-4o cutoff.","analysis":"Post-retirement cleanup is now an operating-control problem instead of a migration reminder, so teams need named owners for temporary transition access, regression checks, and fallback closure before legacy usage quietly extends beyond the intended bridge window.","themes":["deployment","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://help.openai.com/en/articles/11954883-legacy-model-access-for-enterprise-and-edu-users"}]},{"id":"signal-research-2026-04-03-openai-says-gpt-4o-inside-business-enterprise-and-edu-custom-gpts-fully-retires-after-april-3","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-03-openai-says-gpt-4o-inside-business-enterprise-and-edu-custom-gpts-fully-retires-after-april-3","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-03","tag":null,"group":"Vendor governance hardened at the legal and privacy layer","title":"OpenAI says GPT-4o inside Business, Enterprise, and Edu Custom GPTs fully retires after April 3.","analysis":"Teams with Custom GPT workflows tuned to GPT-4o now have a hard migration and QA deadline: validate replacement-model behavior, tool compatibility, and output-schema drift before the cutoff turns model lifecycle management into a same-week continuity issue.","themes":["deployment","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://help.openai.com/en/articles/12003714-chatgpt-business-models-limits"}]},{"id":"signal-research-2026-04-02-openai-launched-flexible-codex-pricing-for-teams-with-a-codex-only-seat-option-and-pay-as-you-go-usage","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-02-openai-launched-flexible-codex-pricing-for-teams-with-a-codex-only-seat-option-and-pay-as-you-go-usage","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-02","tag":null,"group":"Control infrastructure kept moving","title":"OpenAI launched flexible Codex pricing for teams with a Codex-only seat option and pay-as-you-go usage.","analysis":"Agent rollout economics just moved from a simple seat decision to an operating-control issue, so teams need workflow-level spend reporting, approval thresholds, and seat-selection rules before pilot usage starts hiding real cost drift.","themes":["deployment","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/codex-flexible-pricing-for-teams/"}]},{"id":"signal-research-2026-04-02-openai-rolled-out-chatgpt-in-apple-carplay-with-voice-first-access-and-explicit-product-limits","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-04-02-openai-rolled-out-chatgpt-in-apple-carplay-with-voice-first-access-and-explicit-product-limits","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-04-02","tag":null,"group":"Control infrastructure kept moving","title":"OpenAI rolled out ChatGPT in Apple CarPlay with voice-first access and explicit product limits.","analysis":"A new in-car workflow surface is now live, which means acceptable-use rules, hands-free interaction guidance, and exception tracking need to be treated as operating controls before employees start normalizing AI use in motion.","themes":["deployment","validation","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://help.openai.com/en/articles/20001153"}]},{"id":"signal-research-2026-03-31-openai-pricing-now-flags-hosted-container-billing-per-20-minute-session-starting-march-31","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-31-openai-pricing-now-flags-hosted-container-billing-per-20-minute-session-starting-march-31","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-31","tag":null,"group":"Vendor governance hardened at the legal and privacy layer","title":"OpenAI pricing now flags hosted-container billing per 20-minute session starting March 31.","analysis":"Teams using hosted execution need to update runtime guardrails, budget thresholds, and rollback assumptions before month-end so a pricing-model change does not become silent cost drift in production workflows.","themes":["deployment","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://developers.openai.com/api/docs/pricing"}]},{"id":"signal-documented-2026-03-27-microsoft-and-victoria-university-launched-a-datacentre-academy-tied-directly-to-infrastructure-hiring-demand","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-27-microsoft-and-victoria-university-launched-a-datacentre-academy-tied-directly-to-infrastructure-hiring-demand","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-27","tag":"EP013","group":"Career infrastructure became a frontline operating question","title":"Microsoft and Victoria University launched a Datacentre Academy tied directly to infrastructure hiring demand.","analysis":"Workforce response is starting to move from abstract reskilling language into named training pipelines connected to real role outcomes, salary bands, and employer need.","themes":["ecosystem","deployment"],"sources":[{"type":"published-episode","label":"EP013: Career Infrastructure Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep013-runtime-cost-check"}]},{"id":"signal-research-2026-03-27-chatgpt-updated-box-notion-linear-and-dropbox-apps-with-additional-actions-including-write-capabilities-where-supported","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-27-chatgpt-updated-box-notion-linear-and-dropbox-apps-with-additional-actions-including-write-capabilities-where-supported","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-27","tag":null,"group":"Control infrastructure kept moving","title":"ChatGPT updated Box, Notion, Linear, and Dropbox apps with additional actions, including write capabilities where supported.","analysis":"Connector surfaces are turning into live workflow control points: teams need approval boundaries, audit expectations, and reconnection checks when a chat surface gains the ability to modify records instead of only reading context.","themes":["deployment","validation","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://help.openai.com/en/articles/6825453-chatgpt-release-notes"}]},{"id":"signal-research-2026-03-25-openai-documented-how-the-model-spec-should-drive-behavior-rules-eval-gates-and-ongoing-public-feedback","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-25-openai-documented-how-the-model-spec-should-drive-behavior-rules-eval-gates-and-ongoing-public-feedback","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-25","tag":null,"group":"Vendor governance hardened at the legal and privacy layer","title":"OpenAI documented how the Model Spec should drive behavior rules, eval gates, and ongoing public feedback.","analysis":"Behavior governance is turning into a maintained operating artifact instead of a hidden prompt habit, so teams need explicit clauses, test cases, and release checks that tie acceptable behavior to named review owners.","themes":["governance","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/our-approach-to-the-model-spec/"}]},{"id":"signal-research-2026-03-25-openai-launched-a-safety-bug-bounty-with-a-public-abuse-reporting-intake-path","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-25-openai-launched-a-safety-bug-bounty-with-a-public-abuse-reporting-intake-path","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-25","tag":null,"group":"Control infrastructure kept moving","title":"OpenAI launched a Safety Bug Bounty with a public abuse-reporting intake path.","analysis":"External safety reporting is becoming part of live AI operations: teams need named triage owners, reproducibility standards, and mitigation paths ready before outside abuse reports arrive faster than internal review loops can absorb them.","themes":["security","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/safety-bug-bounty/"}]},{"id":"signal-documented-2026-03-24-openai-upgraded-shopping-in-chatgpt-and-expanded-the-agentic-commerce-protocol-into-product-discovery","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-24-openai-upgraded-shopping-in-chatgpt-and-expanded-the-agentic-commerce-protocol-into-product-discovery","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-24","tag":"EP014","group":"Commerce discovery is moving upstream of the website","title":"OpenAI upgraded shopping in ChatGPT and expanded the Agentic Commerce Protocol into product discovery.","analysis":"The first serious comparison can now happen before a shopper reaches the storefront. That shifts the operator question from presence alone to data quality, category readiness, and what the new surface is really doing.","themes":["ecosystem","deployment"],"sources":[{"type":"published-episode","label":"EP014: Commerce Surface Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep014-commerce-surface-check"}]},{"id":"signal-documented-2026-03-24-sephora-launched-an-app-in-chatgpt-that-ties-recommendations-to-loyalty-and-future-in-app-checkout","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-24-sephora-launched-an-app-in-chatgpt-that-ties-recommendations-to-loyalty-and-future-in-app-checkout","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-24","tag":"EP014","group":"Commerce discovery is moving upstream of the website","title":"Sephora launched an app in ChatGPT that ties recommendations to loyalty and future in-app checkout.","analysis":"The pattern is moving from product presence to workflow design. Discovery, narrowing, and eventual conversion are starting to connect inside the same AI-led experience.","themes":["deployment","ecosystem"],"sources":[{"type":"published-episode","label":"EP014: Commerce Surface Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep014-commerce-surface-check"}]},{"id":"signal-documented-2026-03-24-shopify-said-millions-of-merchants-can-now-sell-in-ai-chats-with-referral-attribution-and-merchant-of-record-control","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-24-shopify-said-millions-of-merchants-can-now-sell-in-ai-chats-with-referral-attribution-and-merchant-of-record-control","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-24","tag":"EP014","group":"Commerce discovery is moving upstream of the website","title":"Shopify said millions of merchants can now sell in AI chats with referral attribution and merchant-of-record control.","analysis":"This is one of the clearest signals yet that AI-driven discovery may become measurable enough to manage like an operating surface instead of a novelty traffic story.","themes":["deployment","validation"],"sources":[{"type":"published-episode","label":"EP014: Commerce Surface Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep014-commerce-surface-check"}]},{"id":"signal-research-2026-03-24-openai-released-teen-safety-policy-prompts-for-gpt-oss-safeguard","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-24-openai-released-teen-safety-policy-prompts-for-gpt-oss-safeguard","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-24","tag":null,"group":"Vendor governance hardened at the legal and privacy layer","title":"OpenAI released teen-safety policy prompts for gpt-oss-safeguard.","analysis":"Youth-facing safety is becoming an operational policy surface instead of a general trust-and-safety aspiration, so teams need age-specific rules, review owners, and escalation paths that can be tested before deployment reaches younger users.","themes":["governance","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/teen-safety-policies-gpt-oss-safeguard/"}]},{"id":"signal-documented-2026-03-20-workspace-analytics-added-analytics-viewer-access-admin-created-surveys-and-impact-survey-timing-updates","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-20-workspace-analytics-added-analytics-viewer-access-admin-created-surveys-and-impact-survey-timing-updates","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-20","tag":"EP012","group":"Adoption visibility became a management discipline","title":"Workspace analytics added analytics viewer access, admin-created surveys, and impact-survey timing updates.","analysis":"Adoption visibility is becoming its own management layer. Usage counts can point to direction, but leaders still need workflow-level evidence before claiming meaningful impact.","themes":["validation"],"sources":[{"type":"published-episode","label":"EP012: Work Visibility Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep012-work-visibility-check"}]},{"id":"signal-research-2026-03-18-nist-and-gsa-partnered-on-ai-evaluation-science-for-federal-procurement","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-18-nist-and-gsa-partnered-on-ai-evaluation-science-for-federal-procurement","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-18","tag":null,"group":"Policy and geographic footprint kept expanding","title":"NIST and GSA partnered on AI evaluation science for federal procurement.","analysis":"Procurement evidence expectations are hardening: teams should name who selects benchmarks, who interprets results, and what thresholds must pass before purchase or expansion decisions move forward.","themes":["governance","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.gsa.gov/about-us/newsroom/news-releases/gsa-and-nist-partner-to-boost-ai-evaluation-science-in-federal-procurement-03182026"}]},{"id":"signal-documented-2026-03-17-openai-said-americans-are-sending-nearly-3-million-compensation-related-messages-per-day-to-chatgpt","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-17-openai-said-americans-are-sending-nearly-3-million-compensation-related-messages-per-day-to-chatgpt","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-17","tag":"EP013","group":"Career infrastructure became a frontline operating question","title":"OpenAI said Americans are sending nearly 3 million compensation-related messages per day to ChatGPT.","analysis":"Workers are already using AI to navigate pay, role expectations, and career decisions. That makes guidance quality and information asymmetry a current operating issue, not a future-of-work thought experiment.","themes":["ecosystem","validation"],"sources":[{"type":"published-episode","label":"EP013: Career Infrastructure Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep013-runtime-cost-check"}]},{"id":"signal-documented-2026-03-13-openai-write-actions-pushed-chatgpt-deeper-into-docs-sheets-outlook-and-calendar-workflows","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-13-openai-write-actions-pushed-chatgpt-deeper-into-docs-sheets-outlook-and-calendar-workflows","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-13","tag":"EP012","group":"Adoption visibility became a management discipline","title":"OpenAI write actions pushed ChatGPT deeper into Docs, Sheets, Outlook, and calendar workflows.","analysis":"Once AI can write into live work surfaces, activation stops being the useful metric. Teams need to know whether the workflow itself improved, slowed down, or quietly shifted into manual repair.","themes":["deployment","validation"],"sources":[{"type":"published-episode","label":"EP012: Work Visibility Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep012-work-visibility-check"}]},{"id":"signal-documented-2026-03-12-openai-posted-a-legal-notice-on-unauthorized-equity-transactions","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-12-openai-posted-a-legal-notice-on-unauthorized-equity-transactions","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-12","tag":"EP009","group":"Governance moved from principles to operating standards","title":"OpenAI posted a legal notice on unauthorized equity transactions.","analysis":"Control teams now need claim-confidence checks in intake so uncertain external claims are handled as risk, not reassurance.","themes":["governance","validation"],"sources":[{"type":"published-episode","label":"EP009: Control Hardening Week","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep009-control-hardening-weekly-ops"}]},{"id":"signal-research-2026-03-12-anthropic-committed-100m-to-the-claude-partner-network","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-12-anthropic-committed-100m-to-the-claude-partner-network","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-12","tag":null,"group":"Enterprise operating models thickened","title":"Anthropic committed $100M to the Claude Partner Network.","analysis":"Implementation-partner quality is becoming a governance variable: teams now need to diligence who delivers the rollout, how control evidence is surfaced, and who owns escalation when partner-led deployments drift.","themes":["governance","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.anthropic.com/news/claude-partner-network"}]},{"id":"signal-research-2026-03-12-infinite-dial-2026-confirmed-record-podcast-reach-and-stronger-cross-format-listening-behavior","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-12-infinite-dial-2026-confirmed-record-podcast-reach-and-stronger-cross-format-listening-behavior","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-12","tag":null,"group":"Creator and distribution workflows started hardening","title":"Infinite Dial 2026 confirmed record podcast reach and stronger cross-format listening behavior.","analysis":"Audience expansion is now strong enough to affect packaging decisions: teams can justify audio-plus-video distribution, but they still need explicit ownership for measurement, channel mix, and rights-safe promotion instead of treating audience growth as automatic monetization proof.","themes":["deployment","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.edisonresearch.com/the-infinite-dial-2026/"}]},{"id":"signal-research-2026-03-12-openai-posted-an-unauthorized-equity-transactions-legal-notice","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-12-openai-posted-an-unauthorized-equity-transactions-legal-notice","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-12","tag":null,"group":"Vendor governance hardened at the legal and privacy layer","title":"OpenAI posted an unauthorized equity-transactions legal notice.","analysis":"Vendor diligence now includes checking whether ownership, transfer, and financing narratives rest on instruments the company says it has not authorized.","themes":["governance","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/policies/unauthorized-openai-equity-transactions/"}]},{"id":"signal-research-2026-03-12-openai-published-2025-california-privacy-rights-request-metrics","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-12-openai-published-2025-california-privacy-rights-request-metrics","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-12","tag":null,"group":"Vendor governance hardened at the legal and privacy layer","title":"OpenAI published 2025 California privacy-rights request metrics.","analysis":"Privacy operations are becoming measurable governance signals that teams can benchmark against their own request-handling SLAs and reporting cadence.","themes":["governance","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/2025-california-privacy-rights-request-metrics/"}]},{"id":"signal-research-2026-03-12-rebel-audio-launched-an-ai-native-podcast-platform-and-announced-funding","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-12-rebel-audio-launched-an-ai-native-podcast-platform-and-announced-funding","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-12","tag":null,"group":"Creator and distribution workflows started hardening","title":"Rebel Audio launched an AI-native podcast platform and announced funding.","analysis":"AI-assisted audio production is becoming a workflow-control issue: teams need explicit voice-rights inventories, disclosure rules, and monetization fallback plans before AI-native production stacks become part of the publishing path.","themes":["deployment","security","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.prnewswire.com/news-releases/introducing-rebel-audio-podcasting-for-the-next-generation-of-creators-302711132.html"}]},{"id":"signal-documented-2026-03-11-anthropic-created-the-anthropic-institute","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-11-anthropic-created-the-anthropic-institute","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-11","tag":"EP008 / EP009","group":"Governance moved from principles to operating standards","title":"Anthropic created the Anthropic Institute.","analysis":"Frontier model governance is becoming a durable operating function, not a one-off launch appendix.","themes":["governance"],"sources":[{"type":"published-episode","label":"EP008: AI Brief | EP008: Model release control validation","url":"https://michaelhbm.com/AiChangeDesk/episodes/brief-2026-03-11-ai-brief"},{"type":"published-episode","label":"EP009: Control Hardening Week","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep009-control-hardening-weekly-ops"}]},{"id":"signal-documented-2026-03-11-aws-published-an-agentic-ai-stakeholder-guide-centered-on-named-control-ownership","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-11-aws-published-an-agentic-ai-stakeholder-guide-centered-on-named-control-ownership","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-11","tag":"EP010","group":"Governance moved from principles to operating standards","title":"AWS published an agentic-AI stakeholder guide centered on named control ownership.","analysis":"Agent rollouts now come with an explicit expectation that product, engineering, security, and operations owners define readiness, approvals, and monitoring responsibilities before scale.","themes":["governance","validation","deployment"],"sources":[{"type":"published-episode","label":"EP010: AI Brief | EP010: Evaluation and Ownership Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/brief-2026-03-18-ai-brief"}]},{"id":"signal-documented-2026-03-11-nist-s-monitoring-report-made-deployed-system-evidence-a-frontline-control","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-11-nist-s-monitoring-report-made-deployed-system-evidence-a-frontline-control","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-11","tag":"EP007 / EP008 / EP009","group":"Security workflow controls had to become explicit","title":"NIST’s monitoring report made deployed-system evidence a frontline control.","analysis":"If the action chain is not reconstructable, expansion should pause until the evidence path is fixed.","themes":["security","validation"],"sources":[{"type":"published-episode","label":"EP007: Security Workflow Control Contract","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep007-security-workflow-control-contract"},{"type":"published-episode","label":"EP008: AI Brief | EP008: Model release control validation","url":"https://michaelhbm.com/AiChangeDesk/episodes/brief-2026-03-11-ai-brief"},{"type":"published-episode","label":"EP009: Control Hardening Week","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep009-control-hardening-weekly-ops"}]},{"id":"signal-documented-2026-03-11-wayfair-reported-2-5-million-corrected-product-tags-and-41-000-automated-tickets-per-month","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-11-wayfair-reported-2-5-million-corrected-product-tags-and-41-000-automated-tickets-per-month","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-11","tag":"EP012","group":"Adoption visibility became a management discipline","title":"Wayfair reported 2.5 million corrected product tags and 41,000 automated tickets per month.","analysis":"The useful pattern is not vendor theater. It is one workflow, one measurable change, and one business effect that somebody can defend without hand-waving.","themes":["deployment","validation"],"sources":[{"type":"published-episode","label":"EP012: Work Visibility Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep012-work-visibility-check"}]},{"id":"signal-research-2026-03-11-google-completed-its-wiz-acquisition-and-positioned-it-as-a-unified-multicloud-security-platform","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-11-google-completed-its-wiz-acquisition-and-positioned-it-as-a-unified-multicloud-security-platform","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-11","tag":null,"group":"Enterprise operating models thickened","title":"Google completed its Wiz acquisition and positioned it as a unified multicloud security platform.","analysis":"Cloud and AI security are consolidating into a bundled control layer, so procurement and architecture reviews now need to weigh security operating-model lock-in, multicloud policy consistency, and automation posture alongside model capability.","themes":["security","deployment","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://blog.google/innovation-and-ai/infrastructure-and-cloud/google-cloud/wiz-acquisition/"}]},{"id":"signal-research-2026-03-11-libsyn-reported-record-creator-payouts-while-podnews-flagged-ad-concentration","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-11-libsyn-reported-record-creator-payouts-while-podnews-flagged-ad-concentration","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-11","tag":null,"group":"Enterprise operating models thickened","title":"Libsyn reported record creator payouts while Podnews flagged ad concentration.","analysis":"Distribution economics are becoming a control signal: concentration and monetization dependencies should be tracked alongside technical rollout risk.","themes":["ecosystem","governance"],"sources":[{"type":"public-source","label":"Public source","url":"https://podnews.net/update/libsyn-ad-revenue-48"}]},{"id":"signal-research-2026-03-11-youtube-documented-the-town-s-audio-to-video-expansion-path-across-youtube-and-spotify","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-11-youtube-documented-the-town-s-audio-to-video-expansion-path-across-youtube-and-spotify","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-11","tag":null,"group":"Creator and distribution workflows started hardening","title":"YouTube documented The Town's audio-to-video expansion path across YouTube and Spotify.","analysis":"Video migration is now concrete workflow evidence, not just platform encouragement: teams can study a live cross-format publishing pattern, but they still need named owners for edit cadence, clips, rights-safe reuse, and whether the added video workload weakens the audio release contract.","themes":["deployment","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://blog.youtube/creator-and-artist-stories/matt-belloni-brings-the-town-to-youtube/"}]},{"id":"signal-documented-2026-03-10-google-pushed-gemini-deeper-into-docs-sheets-slides-and-drive","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-10-google-pushed-gemini-deeper-into-docs-sheets-slides-and-drive","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-10","tag":"EP011","group":"AI is disappearing into the default work surface","title":"Google pushed Gemini deeper into Docs, Sheets, Slides, and Drive.","analysis":"The management problem is shifting from tool adoption to workflow friction. AI now lands inside everyday files and collaboration surfaces before many teams update review habits or approval expectations.","themes":["deployment","ecosystem"],"sources":[{"type":"published-episode","label":"EP011: Control Surface Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep011-control-surface-check"}]},{"id":"signal-documented-2026-03-10-podbean-switched-off-dynamic-ad-insertion-in-eea-eu-and-uk-regions","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-10-podbean-switched-off-dynamic-ad-insertion-in-eea-eu-and-uk-regions","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-10","tag":"EP009","group":"Deployment and continuity became executive decisions","title":"Podbean switched off dynamic ad insertion in EEA, EU, and UK regions.","analysis":"Distribution policy shifts can break continuity assumptions just like infrastructure incidents, so fallback ownership has to include platform dependencies.","themes":["deployment","security"],"sources":[{"type":"published-episode","label":"EP009: Control Hardening Week","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep009-control-hardening-weekly-ops"}]},{"id":"signal-documented-2026-03-10-youtube-expanded-likeness-detection-protections-to-journalists-and-public-officials","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-10-youtube-expanded-likeness-detection-protections-to-journalists-and-public-officials","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-10","tag":"EP010","group":"Security workflow controls had to become explicit","title":"YouTube expanded likeness-detection protections to journalists and public officials.","analysis":"Identity-integrity response now needs named triage, verification, pause authority, and public-response ownership before external channels are exposed.","themes":["governance","security","ecosystem"],"sources":[{"type":"published-episode","label":"EP010: AI Brief | EP010: Evaluation and Ownership Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/brief-2026-03-18-ai-brief"}]},{"id":"signal-documented-2026-03-09-codex-security-made-ai-assisted-security-operations-a-workflow-design-problem","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-09-codex-security-made-ai-assisted-security-operations-a-workflow-design-problem","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-09","tag":"EP007","group":"Security workflow controls had to become explicit","title":"Codex Security made AI-assisted security operations a workflow design problem.","analysis":"If AI can read code, propose changes, or open a PR, then ownership, merge authority, and evidence requirements have to be explicit before the tool touches production-bound workflows.","themes":["security","validation"],"sources":[{"type":"published-episode","label":"EP007: Security Workflow Control Contract","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep007-security-workflow-control-contract"}]},{"id":"signal-documented-2026-03-09-gpt-5-4-and-chatgpt-for-excel-showed-how-execution-surfaces-can-expand-inside-familiar-tools","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-09-gpt-5-4-and-chatgpt-for-excel-showed-how-execution-surfaces-can-expand-inside-familiar-tools","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-09","tag":"EP007 / EP008","group":"Deployment and continuity became executive decisions","title":"GPT-5.4 and ChatGPT for Excel showed how execution surfaces can expand inside familiar tools.","analysis":"Teams often treat same-app new-feature changes as low-friction, but behavior changes inside common interfaces can quietly widen approval burden and operator training requirements.","themes":["deployment","access","validation"],"sources":[{"type":"published-episode","label":"EP007: Security Workflow Control Contract","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep007-security-workflow-control-contract"},{"type":"published-episode","label":"EP008: AI Brief | EP008: Model release control validation","url":"https://michaelhbm.com/AiChangeDesk/episodes/brief-2026-03-11-ai-brief"}]},{"id":"signal-documented-2026-03-09-microsoft-s-wave-3-push-embedded-copilot-deeper-inside-word-excel-powerpoint-and-outlook","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-09-microsoft-s-wave-3-push-embedded-copilot-deeper-inside-word-excel-powerpoint-and-outlook","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-09","tag":"EP012","group":"Adoption visibility became a management discipline","title":"Microsoft's Wave 3 push embedded Copilot deeper inside Word, Excel, PowerPoint, and Outlook.","analysis":"The question is no longer whether AI is arriving as a separate tool. The question is whether managers can see where it changed the work once it became part of the default surface.","themes":["deployment","ecosystem"],"sources":[{"type":"published-episode","label":"EP012: Work Visibility Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep012-work-visibility-check"}]},{"id":"signal-documented-2026-03-09-mozilla-and-anthropic-tied-red-teaming-and-coordinated-disclosure-to-workflow-maturity","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-09-mozilla-and-anthropic-tied-red-teaming-and-coordinated-disclosure-to-workflow-maturity","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-09","tag":"EP007","group":"Security workflow controls had to become explicit","title":"Mozilla and Anthropic tied red teaming and coordinated disclosure to workflow maturity.","analysis":"Discovery throughput is not remediation readiness. Someone has to own triage, approval, patch path, user communication, and rollback if the loop finds something real.","themes":["security"],"sources":[{"type":"published-episode","label":"EP007: Security Workflow Control Contract","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep007-security-workflow-control-contract"}]},{"id":"signal-documented-2026-03-09-openai-s-planned-promptfoo-acquisition-signaled-that-evaluation-tooling-is-becoming-core-infrastructure","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-09-openai-s-planned-promptfoo-acquisition-signaled-that-evaluation-tooling-is-becoming-core-infrastructure","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-09","tag":"EP008","group":"Release validation started moving into the platform stack","title":"OpenAI’s planned Promptfoo acquisition signaled that evaluation tooling is becoming core infrastructure.","analysis":"Prompt and agent validation are moving out of nice-to-have red-team craft and into the production release pipeline.","themes":["validation"],"sources":[{"type":"published-episode","label":"EP008: AI Brief | EP008: Model release control validation","url":"https://michaelhbm.com/AiChangeDesk/episodes/brief-2026-03-11-ai-brief"}]},{"id":"signal-research-2026-03-09-anthropic-announced-a-collaboration-with-infosys","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-09-anthropic-announced-a-collaboration-with-infosys","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-09","tag":null,"group":"Enterprise operating models thickened","title":"Anthropic announced a collaboration with Infosys.","analysis":"Partner channels are becoming part of how organizations buy and operationalize model capability.","themes":["ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.anthropic.com/news/anthropic-infosys"}]},{"id":"signal-research-2026-03-09-microsoft-introduced-frontier-suite-and-a-new-transformation-packaging-layer","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-09-microsoft-introduced-frontier-suite-and-a-new-transformation-packaging-layer","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-09","tag":null,"group":"Enterprise operating models thickened","title":"Microsoft introduced Frontier Suite and a new transformation packaging layer.","analysis":"Large vendors are bundling trust, copilots, and services into a more opinionated operating model.","themes":["ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://blogs.microsoft.com/blog/2026/03/09/introducing-the-first-frontier-suite-built-on-intelligence-trust/"}]},{"id":"signal-research-2026-03-09-microsoft-outlined-a-secure-agentic-operating-model-for-frontier-transformation","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-09-microsoft-outlined-a-secure-agentic-operating-model-for-frontier-transformation","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-09","tag":null,"group":"Enterprise operating models thickened","title":"Microsoft outlined a secure-agentic operating model for Frontier transformation.","analysis":"The suite story now includes explicit agent identity, runtime isolation, telemetry, and approval-path requirements instead of leaving security to downstream implementation.","themes":["security","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.microsoft.com/en-us/security/blog/2026/03/09/secure-agentic-ai-for-your-frontier-transformation/"}]},{"id":"signal-research-2026-03-09-microsoft-previewed-a-security-dashboard-for-ai","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-09-microsoft-previewed-a-security-dashboard-for-ai","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-09","tag":null,"group":"Control infrastructure kept moving","title":"Microsoft previewed a Security Dashboard for AI.","analysis":"Monitoring and posture views are becoming first-class management surfaces for deployed AI.","themes":["security","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://techcommunity.microsoft.com/blog/microsoft-security-blog/introducing-security-dashboard-for-ai-now-in-public-preview/4494637"}]},{"id":"signal-research-2026-03-09-openai-and-tata-announced-a-strategic-collaboration","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-09-openai-and-tata-announced-a-strategic-collaboration","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-09","tag":null,"group":"Enterprise operating models thickened","title":"OpenAI and Tata announced a strategic collaboration.","analysis":"Regional delivery and implementation layers are becoming part of the AI operating stack.","themes":["ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.tata.com/newsroom/business/tata-openai-partnership"}]},{"id":"signal-research-2026-03-06-openai-launched-openai-for-india","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-06-openai-launched-openai-for-india","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-06","tag":null,"group":"Policy and geographic footprint kept expanding","title":"OpenAI launched OpenAI for India.","analysis":"Localized delivery models are becoming a strategic part of how major vendors scale adoption.","themes":["ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/openai-for-india"}]},{"id":"signal-documented-2026-03-05-anthropic-expanded-in-sydney-as-its-fourth-asia-pacific-office","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-05-anthropic-expanded-in-sydney-as-its-fourth-asia-pacific-office","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-05","tag":"EP010","group":"Governance moved from principles to operating standards","title":"Anthropic expanded in Sydney as its fourth Asia-Pacific office.","analysis":"Regional operating footprint increasingly matters for support coverage, delivery models, and partner strategy.","themes":["ecosystem","governance"],"sources":[{"type":"published-episode","label":"EP010: AI Brief | EP010: Evaluation and Ownership Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/brief-2026-03-18-ai-brief"}]},{"id":"signal-documented-2026-03-05-openai-argued-that-ai-use-in-education-is-becoming-an-opportunity-gap-not-just-a-classroom-tooling-choice","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-05-openai-argued-that-ai-use-in-education-is-becoming-an-opportunity-gap-not-just-a-classroom-tooling-choice","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-05","tag":"EP013","group":"Career infrastructure became a frontline operating question","title":"OpenAI argued that AI use in education is becoming an opportunity gap, not just a classroom tooling choice.","analysis":"Education is turning into labor-market infrastructure. The management question is no longer whether students can touch AI, but whether access, practice, and guidance arrive early enough to matter before the workplace has already moved.","themes":["ecosystem","deployment"],"sources":[{"type":"published-episode","label":"EP013: Career Infrastructure Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep013-runtime-cost-check"}]},{"id":"signal-documented-2026-03-05-openai-positioned-gpt-5-4-for-professional-work-rather-than-generic-capability-demos","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-05-openai-positioned-gpt-5-4-for-professional-work-rather-than-generic-capability-demos","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-05","tag":"EP011","group":"AI is disappearing into the default work surface","title":"OpenAI positioned GPT-5.4 for professional work rather than generic capability demos.","analysis":"Model choice is turning into an error-budget and review-burden decision. Teams increasingly need to separate the fastest model from the model they trust most in the workflow that matters.","themes":["validation","deployment"],"sources":[{"type":"published-episode","label":"EP011: Control Surface Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep011-control-surface-check"}]},{"id":"signal-documented-2026-03-04-gpt-5-3-instant-raised-the-bar-for-model-continuity-controls","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-03-04-gpt-5-3-instant-raised-the-bar-for-model-continuity-controls","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-04","tag":"EP006","group":"Deployment and continuity became executive decisions","title":"GPT-5.3 Instant raised the bar for model continuity controls.","analysis":"Every material model swap needs regression prompts, fallback path, and a named rollback owner before broader rollout.","themes":["deployment","validation"],"sources":[{"type":"published-episode","label":"EP006: AI Brief: GPT-5.3 and continuity controls","url":"https://michaelhbm.com/AiChangeDesk/episodes/brief-2026-03-04-ai-brief"}]},{"id":"signal-research-2026-03-04-aws-added-stateful-mcp-support-to-bedrock-agentcore-runtime","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-04-aws-added-stateful-mcp-support-to-bedrock-agentcore-runtime","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-04","tag":null,"group":"Control infrastructure kept moving","title":"AWS added stateful MCP support to Bedrock AgentCore runtime.","analysis":"State management and multi-tool coordination are becoming explicit platform capabilities, not just custom engineering work.","themes":["deployment","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://aws.amazon.com/about-aws/whats-new/2026/03/amazon-bedrock-agentcore-runtime-stateful-mcp/"}]},{"id":"signal-research-2026-03-02-openai-updated-its-department-of-war-agreement-with-explicit-domestic-surveillance-and-nsa-limits","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-03-02-openai-updated-its-department-of-war-agreement-with-explicit-domestic-surveillance-and-nsa-limits","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-03-02","tag":null,"group":"Vendor governance hardened at the legal and privacy layer","title":"OpenAI updated its Department of War agreement with explicit domestic-surveillance and NSA limits.","analysis":"Vendor diligence now includes checking whether sensitive-government deployments rely on cloud-only architecture, human oversight, and contract language that keeps surveillance and intelligence-use boundaries enforceable even if policy pressure changes later.","themes":["governance","deployment","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/our-agreement-with-the-department-of-war/"}]},{"id":"signal-research-2026-02-28-aws-added-server-side-tool-execution-to-bedrock-agentcore","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-02-28-aws-added-server-side-tool-execution-to-bedrock-agentcore","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-28","tag":null,"group":"Control infrastructure kept moving","title":"AWS added server-side tool execution to Bedrock AgentCore.","analysis":"Tooling and execution control are moving deeper into managed infrastructure layers.","themes":["deployment","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://aws.amazon.com/about-aws/whats-new/2026/02/amazon-bedrock-server-side-tool-execution-agentcore-gateway/"}]},{"id":"signal-documented-2026-02-25-openai-introduced-lockdown-mode-and-elevated-risk-labels-in-chatgpt-safety","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-02-25-openai-introduced-lockdown-mode-and-elevated-risk-labels-in-chatgpt-safety","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-25","tag":"EP005","group":"Access control became the real agent risk","title":"OpenAI introduced Lockdown Mode and elevated risk labels in ChatGPT safety.","analysis":"App-level approval stopped being granular enough. Different action surfaces inside the same interface now require different control states.","themes":["access","security"],"sources":[{"type":"published-episode","label":"EP005: Run Agents Without Losing Control","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep005-access-controls-operators"}]},{"id":"signal-documented-2026-02-25-openai-s-misuse-report-reinforced-the-need-for-evidence-and-stop-authority","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-02-25-openai-s-misuse-report-reinforced-the-need-for-evidence-and-stop-authority","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-25","tag":"EP005","group":"Access control became the real agent risk","title":"OpenAI’s misuse report reinforced the need for evidence and stop authority.","analysis":"Without reconstructable logs and named pause authority, incident handling becomes storytelling instead of response.","themes":["access","security"],"sources":[{"type":"published-episode","label":"EP005: Run Agents Without Losing Control","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep005-access-controls-operators"}]},{"id":"signal-documented-2026-02-24-eu-proportionality-guidance-made-risk-scaling-more-concrete","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-02-24-eu-proportionality-guidance-made-risk-scaling-more-concrete","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-24","tag":"EP005","group":"Governance moved from principles to operating standards","title":"EU proportionality guidance made risk scaling more concrete.","analysis":"Connected, hybrid, and disconnected deployments should not share one approval memo. The control set has to scale with deployment type and consequence.","themes":["governance","deployment"],"sources":[{"type":"published-episode","label":"EP005: Run Agents Without Losing Control","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep005-access-controls-operators"}]},{"id":"signal-documented-2026-02-24-microsoft-sovereign-cloud-ai-updates-pushed-deployment-tiering-into-practical-governance","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-02-24-microsoft-sovereign-cloud-ai-updates-pushed-deployment-tiering-into-practical-governance","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-24","tag":"EP005","group":"Deployment and continuity became executive decisions","title":"Microsoft sovereign cloud AI updates pushed deployment tiering into practical governance.","analysis":"Connected, hybrid, and sovereign-style environments carry different key custody, audit, latency, and rollback expectations.","themes":["deployment","governance"],"sources":[{"type":"published-episode","label":"EP005: Run Agents Without Losing Control","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep005-access-controls-operators"}]},{"id":"signal-research-2026-02-24-anthropic-released-responsible-scaling-policy-v3","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-02-24-anthropic-released-responsible-scaling-policy-v3","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-24","tag":null,"group":"Policy and geographic footprint kept expanding","title":"Anthropic released Responsible Scaling Policy v3.","analysis":"Vendor safety documentation is part of the release signal and should not be treated as static background material.","themes":["governance","validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.anthropic.com/news/responsible-scaling-policy-v3"}]},{"id":"signal-documented-2026-02-23-anthropic-s-ai-fluency-index-showed-users-challenge-polished-output-less-once-it-looks-finished","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-02-23-anthropic-s-ai-fluency-index-showed-users-challenge-polished-output-less-once-it-looks-finished","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-23","tag":"EP011","group":"AI is disappearing into the default work surface","title":"Anthropic's AI Fluency Index showed users challenge polished output less once it looks finished.","analysis":"Challenge habits now matter as much as access controls. When output feels clean and complete, teams are more likely to skip the second look that would normally catch missing context or weak reasoning.","themes":["validation"],"sources":[{"type":"published-episode","label":"EP011: Control Surface Check","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep011-control-surface-check"}]},{"id":"signal-research-2026-02-23-openai-launched-frontier-alliances","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-02-23-openai-launched-frontier-alliances","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-23","tag":null,"group":"Enterprise operating models thickened","title":"OpenAI launched Frontier Alliances.","analysis":"Enterprise AI deployment is increasingly being sold as an operating-model redesign, not just a software purchase.","themes":["governance","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://openai.com/index/frontier-alliance-partners"}]},{"id":"signal-documented-2026-02-20-nist-pushed-agent-interoperability-and-efficiency-into-the-standards-workflow","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-02-20-nist-pushed-agent-interoperability-and-efficiency-into-the-standards-workflow","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-20","tag":"EP003 / EP004","group":"Governance moved from principles to operating standards","title":"NIST pushed agent interoperability and efficiency into the standards workflow.","analysis":"Once NIST opened the standards workflow, interoperability stopped being a future talking point and became procurement language teams can start asking for now.","themes":["governance"],"sources":[{"type":"published-episode","label":"EP003: AI governance implementation for operators: turning policy into weekly execution","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep003-ai-governance-implementation"},{"type":"published-episode","label":"EP004: AI Brief: what changed this week","url":"https://michaelhbm.com/AiChangeDesk/episodes/brief-2026-02-25-ai-brief"}]},{"id":"signal-documented-2026-02-18-anthropic-s-distillation-attack-work-widened-the-control-conversation","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-02-18-anthropic-s-distillation-attack-work-widened-the-control-conversation","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-18","tag":"EP004","group":"Access control became the real agent risk","title":"Anthropic’s distillation attack work widened the control conversation.","analysis":"It is no longer only about what a model can do for you. It is also about how capability can be copied, extracted, or bypassed when control boundaries are weak.","themes":["access","security"],"sources":[{"type":"published-episode","label":"EP004: AI Brief: what changed this week","url":"https://michaelhbm.com/AiChangeDesk/episodes/brief-2026-02-25-ai-brief"}]},{"id":"signal-documented-2026-02-17-claude-sonnet-4-6-and-claude-code-security-reinforced-that-capability-and-secure-rollout-now-ship-together","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-documented-2026-02-17-claude-sonnet-4-6-and-claude-code-security-reinforced-that-capability-and-secure-rollout-now-ship-together","lane":"documented","status":"Operating implication covered in a published AI Change Desk episode file","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-17","tag":"EP003","group":"Release validation started moving into the platform stack","title":"Claude Sonnet 4.6 and Claude Code security reinforced that capability and secure rollout now ship together.","analysis":"Model progress now arrives alongside stronger expectations about secure use inside developer workflows.","themes":["validation","security"],"sources":[{"type":"published-episode","label":"EP003: AI governance implementation for operators: turning policy into weekly execution","url":"https://michaelhbm.com/AiChangeDesk/episodes/ep003-ai-governance-implementation"}]},{"id":"signal-research-2026-02-17-caisi-scheduled-listening-sessions-on-ai-adoption-barriers","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-02-17-caisi-scheduled-listening-sessions-on-ai-adoption-barriers","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-17","tag":null,"group":"Policy and geographic footprint kept expanding","title":"CAISI scheduled listening sessions on AI adoption barriers.","analysis":"Adoption barriers are being treated as a standards and policy concern, not only an internal project issue.","themes":["governance","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.nist.gov/news-events/news/2026/02/caisi-host-listening-sessions-barriers-ai-adoption"}]},{"id":"signal-research-2026-02-17-nist-expanded-its-ai-evaluation-toolbox","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-02-17-nist-expanded-its-ai-evaluation-toolbox","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-17","tag":null,"group":"Control infrastructure kept moving","title":"NIST expanded its AI evaluation toolbox.","analysis":"Evaluation maturity is widening beyond benchmark leaderboards into stronger measurement practice and uncertainty handling.","themes":["validation"],"sources":[{"type":"public-source","label":"Public source","url":"https://www.nist.gov/publications/expanding-ai-evaluation-toolbox-statistical-models"}]},{"id":"signal-research-2026-02-16-apple-made-video-podcast-publishing-operational-with-hls-first-workflow-guidance-and-rss-support","url":"https://michaelhbm.com/AiChangeDesk/changes#signal-research-2026-02-16-apple-made-video-podcast-publishing-operational-with-hls-first-workflow-guidance-and-rss-support","lane":"research","status":"Source-linked signal in the editorial research queue","recordKind":"operating-signal","dateBasis":"source-date","date":"2026-02-16","tag":null,"group":"Creator and distribution workflows started hardening","title":"Apple made video podcast publishing operational with HLS-first workflow guidance and RSS support.","analysis":"Podcast teams now need to choose a concrete video architecture, define rollout QA, and separate 'video enabled' from 'monetization ready' before distribution complexity quietly spreads across hosting, ad ops, and playback support.","themes":["deployment","validation","ecosystem"],"sources":[{"type":"public-source","label":"Public source","url":"https://podcasters.apple.com/support/3684-video-podcasts"}]}]}