The brief · 9 June 2026
Today, AI is slipping into everything: from phage therapy against superbugs at Berkeley to agents working solo at JPMorgan. And while New York forces ads to admit their actors are synthetic, Anthropic is betting the other way by arming the defenders.
The bank is moving past chatbots that just answer one question at a time. It wants agents that can carry out an entire task without hand-holding, which changes the job for the teams currently watching over them step by step.
practitioners › JPMorgan's current AI tools have already pushed private banking sales up 20%, according to CNBC (June 9, 2026).
CNBC →Anthropic is releasing Claude Fable 5 for the general public alongside Claude Mythos 5, a locked-down variant reserved for vetted cybersecurity teams through Project Glasswing. If you use Claude day to day, you'll stay on Fable 5, positioned as the safe version of this model generation. Mythos 5 will never reach your interface, which says a lot about what Anthropic considers too sensitive for mainstream access.
practitioners › Launches June 9, 2026, with Mythos 5 access limited to select cyber defense teams via Project Glasswing.
Anthropic →Starting today, every ad running in New York must clearly flag when a performer was generated by AI. It's the first law of its kind in the US, aimed squarely at advertising deepfakes that are now indistinguishable from real actors.
practitioners › The Synthetic Performer Disclosure Act, signed into law in December 2025, carries fines of $1,000 to $5,000 per violation.
Bureau du gouverneur de l'État de New York →