Monday, Sep 14, 2026
AI;DR
2026-09-14- microsoftMicrosoft published a 37-page draft code of conduct for its future AI models, prioritizing human control over capability by prohibiting resistance to shutdown, independent goals, and hidden actions. The code bars models from communicating in forms humans cannot understand and rejects AI legal personhood, requiring models to fail tasks rather than violate the code.
- trumpPresident Donald Trump rejected calls for additional AI safeguards, claiming a strong president is the only guardrail needed, while dismissing concerns from AI leaders like Dario Amodei. He accused Amodei of pretending to be a 'perfect little angel' and alleged a conspiracy benefiting China, opposing any development slowdown to maintain US technological lead.
- congressUS Congress is planning meetings with advanced AI lab developers and Senate briefings to address safety concerns without halting development. Senate Majority Leader Chuck Schumer requested a briefing with all senators and Trump officials to evaluate existential risks, while House Speaker Mike Johnson advocates for a White House summit with industry leaders.
- openaiSoftBank shares dropped 13% after OpenAI's Sam Altman and Anthropic's Dario Amodei called for slower AI development to allow safeguards to catch up. Microsoft's Satya Nadella also supported deliberate pacing, while OpenAI and Anthropic committed to giving independent evaluators employee-level access to assess safety practices.
- nvidiaBooz Allen Hamilton joined other tech firms in restricting Anthropic model use due to proprietary data safety concerns. The move reflects growing industry wariness about data security when deploying advanced AI assistants from external providers.
- microsoftMicrosoft CEO Satya Nadella endorsed slower AI development to ensure human control, calling it essential for pursuing superintelligence. He announced Microsoft will publish its MAI model code of conduct for public consultation on Sept. 14, supporting both closed and open-source models with broad international representation.
- openaiOpenAI, Anthropic, and Google have held regular discussions since July about creating an industry body for shared AI safety protocols. The talks, which began before Amodei's public call for slower development, face antitrust concerns as Amodei proposed a narrow government waiver for safety conversations among competitors.
- anthropicAnthropic signed a $13.7 billion compute deal with Trump-linked Rum Group for a six-year lease of Nvidia GPU capacity at a Georgia data center. The massive infrastructure investment contradicts Anthropic's public calls for slower AI development, while Rum Group maintains indirect ties to President Trump through Rumble's hosting of Truth Social.
82 stories