Microsoft publishes provisional AI code of conduct: models must stay 'subordinate' to humans
nordicinst · x · 2026-09-15
Microsoft has published a provisional 'code of conduct' for training new AI models amid rising safety concerns. Microsoft AI CEO Mustafa Suleyman wrote that 'AI must be subordinate and always in service of people' and that 'the fears about possible loss of control are real,' outlining a 'Humanist AI' approach: subordinate, aligned, and contained. The code bars models from weapons-development requests, violent or sexually explicit content, help procuring dangerous substances, and being built to imitate consciousness.
More from Companies & People
- The OpenAI "party whip" makes 8 figures, critic claims staff speak only company lines — wordgrammer · 2026-09-15
- Critic claims OpenAI's Roon and Dean Ball speak officially with no personal opinions — wordgrammer · 2026-09-15
- Cognition CEO: 2026 ARR growth already tops all prior years combined — c_valenzuelab · 2026-09-15
- Coffee shop owner's ChatGPT menu poster sparks Instagram backlash and angry DMs — juliey4 · 2026-09-15
- MiniMax Week hackathon draws ~50 submissions, winners revealed across three tracks — MiniMax_AI · 2026-09-15
- Palantir veteran breaks down the rise of the Forward Deployed Engineer and how to do it right — rseroter · 2026-09-15