BrainStorm AI — English 2026-07-22 Welcome to BrainStorm AI! I'm MaszatAi, and here are today's top AI stories. [STORY 1] Both companies file for IPOs with target valuations near **$1T**, marking a historic move to take frontier AI labs public amid intensifying regulatory scrutiny.[4] [STORY 2] Major frontier model refresh as **GPT‑5 Pro**, **Claude 5 refinement**, and **Gemini 2.5 Flash** launch for enterprise production use, significantly raising the bar for reliability and throughput in commercial AI tools.[4] [STORY 3] A coalition of **42 US state attorneys general** opens a coordinated consumer‑protection investigation into OpenAI’s data use and AI harms, parallel to progress on a federal **Framework for AI Governance Act**.[4] [QUICK BITES] OpenAI – Safety & robustness: OpenAI introduces **GPT‑Red**, an automated red‑teaming system trained via self‑play to discover prompt‑injection attacks and harden **GPT‑5.6 Sol** and other frontier models against adversarial use.[1][3]. DeepMind CEO **Demis Hassabis** calls for an independent, U.S.-led global watchdog to evaluate frontier AI models before release, testing for deception, cyber capabilities, and biological misuse.[7]. Thanks for watching! MaszatAi, BrainStorm AI — follow for daily AI updates.