Anthropic launches Claude Fable 5 — its most capable model yet with new safety guardrails. Google DeepMind commits $10M to study multi-agent AI risks. NIST provides mathematical proof for continuous AI security monitoring.
OpenAI launches Lockdown Mode to combat prompt injection attacks. Meta's AI support agent exploited to hijack Instagram accounts. NSA reportedly preparing Anthropic's Mythos for cyber operations. Industry grapples with rising AI infrastructure costs.
Brian Chesky launches AI lab, Meta's Business Agent goes global, Trump signs narrower AI executive order, Anthropic files for IPO, and OpenAI Codex hits 5M weekly users.
Anthropic files for IPO in confidential SEC filing, reaching $965B valuation. Microsoft and Google ramp up AI coding tools to compete with Anthropic and OpenAI. EU cybersecurity agency gains access to Mythos AI.