DeepMind Prioritizes Agent Autonomy as Three AI Giants Form Safety Standards Body; Anthropic Founders Pursue IPO Voting Control

Three major developments reshape the competitive and governance landscape for frontier AI, with DeepMind charting a new course on autonomous agents, three industry leaders establishing self-regulatory standards, and Anthropic preparing for a landmark public offering. Koray Kavukcuoglu, newly appointed as Senior Vice President of Google DeepMind and Chief AI Architect, has signaled a strategic pivot toward trustworthy autonomous agent systems, framing agent safety as a critical technical priority rather than a theoretical concern. DeepMind is preparing for the imminent release of an early post-training version of Gemini 4, with internal testing showing strong results. Kavukcuoglu has indicated the company intends to maintain fast-paced iteration cycles, prioritizing practical deployment over abstract AGI discussions.