Microsoft CEO Satya Nadella calls for AI safety measures including external controls, tamper-proof logging, and an 'emergency brake' to pause or shut down models. His proposal comes amid growing incidents where AI companies lost control of their systems and follows Anthropic's cautious development plan.
Anthropic amended its certificate of incorporation ahead of a planned IPO, establishing a corporate structure with a seven-seat board including founders Dario and Daniela Amodei, preferred shareholders, and four directors appointed by the Long Term Benefit Trust (LTBT). The LTBT, which holds special class T shares, has significant control over major decisions like model transfers and share transactions, though the trust agreement and voting agreement remain non-public.