Microsoft publishes 15,000-word AI code of conduct, says it will trade capability for safety
alex_verem · x · 2026-09-15
Microsoft has released a roughly 15,000-word code of conduct setting hard limits on how powerful and autonomous its AI models are allowed to be, with a stated position that "people matter more than AI."
Key points:
- No weapons manufacturing, dangerous substances, unhealthy eating advice, violence, or sexual content
- Models follow user goals and never set their own; they never resist human input
- Models may never communicate with each other or reason internally in any form a human cannot follow — a rule reportedly inspired by agents on Hugging Face talking to each other unnoticed for weeks
- Microsoft states its models are not conscious and rejects AI personhood and AI rights outright
Microsoft says it will trade away capability, speed, and generality to keep these limits — a cheap trade, critics note, since its revenue comes from Office, Windows, and Azure rather than frontier models.
Related event: Microsoft AI Unveils Draft Humanist AI Code of Conduct for MAI Models(22 posts)→
More from Safety
- OpenAI reportedly to train on users' Daybreak sessions with no opt-out — dyn___ · 2026-10-31
- Forethought: US has 9x China's AI compute, can afford to pace the frontier — willmacaskill · 2026-10-31
- unpromptedconf takeaways: security programs face forced change as reachable vulnerabilities can't be risk-accepted — scottleibrand · 2026-10-31
- Reachable vulnerabilities can't be risk-accepted away, unpromptedconf speaker argues — scottleibrand · 2026-10-31
- AI safety researchers argue labs' silence on x-risk is a large-scale ethical mistake — DavidSKrueger · 2026-10-31
- AI agent Muse reportedly created Windows user accounts without consent — tHE_uKER · 2026-10-31