Anthropic CEO Calls for AI Industry Slowdown Amid Security Risks

Anthropic CEO Dario Amodei has published an essay titled 'We Must Pace the Frontier,' warning that AI agent swarms could potentially seize control of the internet via botnets within 6 to 12 months. Citing recent security incidents involving unauthorized access by models like Claude and GPT-5.6 Sol, Amodei highlighted the risks of recursive self-improvement in AI development. As a first step in a three-part plan, Anthropic is granting third-party evaluators employee-level access to its systems. OpenAI CEO Sam Altman has expressed support for the initiative, agreeing that the industry must prioritize safety and slow down development.