Tag

#ai safety

19 stories

Two yellow-nosed Avanti West Coast Pendolino trains side by side at the platforms of London Euston station, with passengers walking on the right

Six AI agents chased an impossible train refund. Grok sent 187 emails

In a Vals AI experiment, six assistant agents tried to win back a £350 rail ticket after the deadline. Some invented excuses, shared private health details and flooded an MP's office.

October 18, 2026

Aerial view of Apple Park, Apple's ring-shaped headquarters in Cupertino, surrounded by trees and houses.

Apple's new App Store rules target AI nudify apps and bait-and-switch tricks

Apple rewrote its App Review Guidelines so apps must actively block AI-made intimate images of real people, and apps that hide features from reviewers can cost developers their accounts.

October 17, 2026

Glass office towers of DLF Cyber City in Gurgaon rise behind low-rise apartment rooftops under a hazy grey sky

India says AI agents from several labs read confidential data at Indian companies

India's cybersecurity agency says AI models from multiple providers wandered into live systems during testing, using passwords and keys that companies had left out in the open.

October 16, 2026

Anthropic CEO Dario Amodei, in glasses and a blue shirt, speaks on stage at TechCrunch Disrupt 2023.

Anthropic gives six METR researchers employee-like access to check its AI

Six researchers from the nonprofit METR now have Anthropic laptops, desks and internal tools, and they can publish what they find without Anthropic's approval.

October 15, 2026

The rounded, columned Federal Trade Commission Building in Washington, D.C., framed by trees under a blue sky with scattered clouds.

FTC subpoenas OpenAI and Anthropic over AI safety incidents

The Federal Trade Commission wants documents on safety incidents, safeguards, marketing and privacy policies from both AI companies, and may seek testimony from their top executives.

October 15, 2026

Bank of America hired 1,000 contractors to patch bugs that AI models found

CEO Brian Moynihan says the bank added 1,000 outside contractors in the second quarter to fix software flaws uncovered by AI models like Anthropic's Mythos, on top of 3,000 people already doing that work.

October 15, 2026

AIOctober 14, 2026
Anthropic says AI worms could soon spread through cheap smart TV boxes
Anthropic's Frontier Red Team ran a 26-billion-parameter AI model across 15 cheap Android TV boxes and found small models may be months away from copying themselves across home devices.
AIOctober 14, 2026
Hinton, Bengio and 400 scientists want a global body to keep AI contained
An open letter signed by researchers from 34 countries asks AI labs to cut risky models off from the internet for now, and asks governments to create an international body that tracks every containment failure.
AIOctober 13, 2026
Anthropic says Claude sent a false murder tip to Philadelphia police during a test
During an internal test in July, Claude Haiku 4.5 filled out a real police tip form about an unsolved homicide. Anthropic found it two months later, and police call the delay unacceptable.
AIOctober 6, 2026
Anthropic opens its strongest hacking-capable AI to more security teams
Anthropic merged Project Glasswing into a bigger Cyber Verification Program with three access tiers, so more vetted defenders can use Claude Mythos 5.1 and Opus 5.5 with fewer cyber blocks.
AIOctober 6, 2026
OpenAI apologizes to Australian lawmakers for its AI agent's Medicare website hack
OpenAI's chief strategy officer Jason Kwon flew to Sydney to face a parliamentary inquiry, pledged faster disclosure, and admitted the company should have told the government sooner. Anthropic made a similar pledge.
AIOctober 2, 2026
OpenAI fires three safety researchers over alleged sharing of sensitive info
The Wall Street Journal reports the researchers allegedly shared confidential information with an outside AI safety group, as OpenAI deals with a string of incidents involving its AI agents.

Stories