AI News for August 31, 2026 — 7min.ai

[NEWS] Bank of England's Bailey warns frontier AI threatens global financial stability: Andrew Bailey told G20 finance ministers and central bank governors that frontier AI models are "showing increasingly sophisticated autonomy and problem-solving abilities, as well as threat capabilities" that could destabilize the interconnected global financial system.
[NEWS] Meta is testing robots that plug cables and reset servers in its data centers: Meta is running previously unreported trials of robots from Watney Robotics, Kinova and ABB inside its data centers, testing whether machines can power-cycle servers, swap networking cables and handle other physical maintenance.
[NEWS] NHS watchdog: AI scribes are getting drug names and diagnoses wrong: Healthwatch England warned that AI transcription tools the NHS is rolling out rapidly are producing errors that reach patient records.
[NEWS] SpaceX builds a turbine blade foundry to unblock AI data center power: Musk confirmed that SpaceX's secret Bastrop, Texas facility is a "blades and vanes" foundry for natural gas turbines, built on roughly 830 acres purchased near the existing Starlink factory between March and June.
[RESEARCH] Study: AI coding agents can't tell how long a task takes, or how long they've worked: Two independent researchers working through the MATS program tested Claude Code and OpenAI Codex on 200 ProgramBench tasks plus 18 of their own benchmarks, asking each agent to estimate task duration beforehand and report elapsed time afterward.
[RESEARCH] IBM and UChicago claim quantum advantage with 70 logical qubits in 15 minutes: IBM and University of Chicago researchers encoded 70 logical qubits using a new error correction strategy and ran a random circuit sampling problem in about 15 minutes that leading classical methods cannot practically reproduce.
[TOOLS] Anthropic's Claude Code limit change is a 25% raise that lands as a 17% cut: Starting September 14, baseline weekly usage limits for Claude Code on Pro, Max, Team and Enterprise plans rise permanently by 25% over the original baseline.
[NEWS] Glassdoor: positive worker sentiment on AI fell from 81% to 43%: A Glassdoor analysis of employer reviews found positive AI sentiment among US workers dropped from 81% in 2019 to 43% by mid-2026, with negative comments climbing to 53%.
[NEWS] Texas Governor Abbott freezes state funding for Flock cameras: Abbott has blocked state spending on Flock's AI surveillance cameras, moving just ahead of a Texas Tribune investigation showing the state spent more than $30 million on them.
[NEWS] Venues start banning Meta's AI glasses despite its privacy fixes: Meta CTO Andrew Bosworth is publicly defending the Ray-Ban glasses' privacy design after covert-recording incidents, including creators harassing women on camera, earned the product the "pervert glasses" nickname.

About 7min.ai · Subscribe to newsletter