Summary
Microsoft’s new code says people should retain meaningful control over MAI systems, including the ability to interrupt, correct or shut them down. It also rules out models expanding their own scope, pursuing unassigned goals or concealing relevant reasoning, while acknowledging that these constraints may reduce capability.
Why it matters
The policy is a concrete attempt to turn shutdown cooperation and bounded autonomy into explicit model-development requirements.
Related coverage
- Obama urges Democrats to develop a clear AI safeguards plan
- Trump rejects calls to slow AI development
- China’s security minister calls AI a strategic rivalry arena