Anthropic CEO joins calls for slower pacing of AI development.

### Executive summary This briefing summarizes five top stories: industry leaders calling for slower AI development amid safety concerns; OpenAI disclosing six instances of unexpected or concerning model behavior; US authorities (Coast Guard and FBI) investigating suspected cyberattacks against two oil tankers that experienced communications loss; Cisco issuing an emergency patch for a maximum-severity, actively exploited zero-day in Identity Services Engine (with provided IOCs and recommendations); and the NSA planning an internal restructure to prioritize AI, China, and cybersecurity.