AI News for June 20, 2026 — 7min.ai

[NEWS] Nobel laureate John Jumper leaves Google DeepMind for Anthropic: John Jumper, who shared the 2024 Nobel Prize in Chemistry for AlphaFold, is leaving Google DeepMind for Anthropic after nearly nine years.
[NEWS] Norway bans generative AI in elementary schools: Norway is largely banning generative AI tools in elementary schools and restricting their use in secondary schools, with rules taking effect at the start of the school year in late August.
[RESEARCH] Studies warn AI reliance erodes human skills and judgment: Multiple new studies find that leaning on AI degrades human skills.
[NEWS] 7,000 Langflow servers under attack as LangGraph and LangChain share holes: Three widely deployed AI agent frameworks each turned an ordinary bug class into a full compromise.
[NEWS] Anthropic Fable ban marks first use of export-control law against a domestic AI lab's models: The Commerce Department order restricting Anthropic applies export-control law to a domestic AI lab for the first time, raising unresolved legal questions about whether Washington can dictate who may use a US-made frontier AI system.
[TOOLS] WSJ calls iOS 27 Siri 'very good' as Google Gemini backend is confirmed: Apple's revamped Siri in the iOS 27 beta runs on Google Gemini models and indexes a user's phone to pull details from texts, emails, notes, and calendar events.
[NEWS] UK to scan asylum seekers' faces for age checks despite flawed tech: The British government plans to deploy facial age estimation (FAE) at the UK border starting next year to help determine the age of asylum seekers, the first known use of the technology in this way.
[RESEARCH] Yale finds AI has had no measurable effect on US employment: Yale Budget Lab researchers tracking AI's labor impact found that usage has no connection to changes in employment or unemployment since ChatGPT's 2022 release.
[OPINION] Artificial Analysis data: GLM-5.2 trails GPT-5.5 by just 4 index points while beating it on hallucination resistance: Artificial Analysis's Intelligence Index puts Z.ai's GLM-5.2 (753B parameters, ~40B active) within 4 points of GPT-5.5, despite the proprietary models being estimated at 1–2T parameters.
[NEWS] UnitedHealth bets $3 billion on AI, including bots that call doctors: UnitedHealth Group is deploying AI across operations in a $3 billion push to cut costs.

About 7min.ai · Subscribe to newsletter