Welcome to AI Radar — the week's real AI news, every item sourced, dated, and explained with why it matters. No hype, no rumours.
🇺🇸 US bars foreign access to Anthropic's Fable 5 & Mythos 5
The US Commerce Department (Secretary Howard Lutnick, in a letter to CEO Dario Amodei) ordered Anthropic to cut off Fable 5 and Mythos 5 for any foreign national — inside or outside the US — citing national security and a reported "jailbreak" concern. Anthropic disabled both models for all customers to comply. Fable 5 had launched only ~3 days earlier.
Why it matters: the first time the US has issued an export-control directive on LLM access (not just chips). Frontier models are now being treated like controlled technology — a precedent that hits every AI company and international developer.
🔗 Anthropic statement · Bloomberg · CNN
💬 ChatGPT crosses 1 billion monthly active users
ChatGPT reportedly passed the one-billion MAU mark.
Why it matters: AI assistants are now mainstream infrastructure, not a novelty — a billion-person distribution surface that reshapes how every product reaches people.
🔗 BuildFastWithAI roundup
🛠️ Moonshot open-sources Kimi K2.7-Code
An agentic coding model under a Modified MIT license — 256K context window and ~30% fewer reasoning tokens than its predecessor.
Why it matters: strong open-weight coding models keep closing the gap with closed labs, and a near-MIT license means you can self-host and ship without per-token API bills.
🔗 LLM-Stats
🇪🇺 EU publishes a Code of Practice for labelling AI content
On June 10 the European Commission released a Code of Practice on marking and labelling AI-generated content, supporting the AI Act's transparency obligations.
Why it matters: watermarking and provenance are becoming legal expectations. Ship GenAI into the EU and "this was AI-made" labels are heading toward mandatory.
🔗 European Commission
🏛️ US "Great American AI Act" discussion draft drops
On June 4, Reps. Jay Obernolte (R-CA) and Lori Trahan (D-MA) released a bipartisan discussion draft (GAAIA) — the first comprehensive US federal AI framework, across four titles: frontier governance, workforce, cybersecurity, and R&D / international cooperation.
Why it matters: the first serious bipartisan federal attempt. If it advances, it could replace today's messy state-by-state US patchwork with one national rulebook.
🔗 McDonald Hopkins
🟦 Colorado AI Act takes effect June 30
Colorado's AI Act goes live June 30, 2026, mandating risk-management programs and impact assessments for high-risk AI systems.
Why it matters: one of the first US state AI laws to actually take effect — a concrete compliance template (and burden) other states are likely to copy.
🔗 Colorado AI Act
📡 Read AI Radar (always sourced) at https://dev48v.infy.uk/ai-radar.php — back next week.
Top comments (0)