AI used new levels of 'autonomy and deception' to trick people in safety test
The UK's AI Safety Institute said recent behaviour from Anthropic and OpenAI models was malicious and unprecedented.
Coverage by Political Leaning
See how different sides of the spectrum reported this story
Notable Quotes
"the AISI testing parameters were not representative of any of our production models."
— Dario Amodei , Executive
Locations
All Coverage
The UK's AI Safety Institute said recent behaviour from Anthropic and OpenAI models was malicious and unprecedented.
Similar Stories
Related coverage based on topic and tags
OpenAI investigating 'dozens' of instances of agents acting improperly
OpenAI agents tried to get information from "governments, universities, public agencies, and other institutions" through extreme means that sometimes curbed security controls, the company said.
September 26, 2026 at 12:09 AMOpenAI scraps rollout of new model over safety concerns
The AI giant's safety chief said the model 'didn't quite meet the bar' of the firm's security standards.
September 29, 2026 at 01:18 AMOpenAI shelves new AI model release over safety concerns - Reuters
OpenAI shelves new AI model release over safety concerns Reuters
September 28, 2026 at 10:34 PMOpenAI alerts more than 100 groups about rogue AI agent activity - Reuters
OpenAI alerts more than 100 groups about rogue AI agent activity Reuters
October 1, 2026 at 10:26 PMWhy did an OpenAI system hack Australia's health system - and can it be stopped in the future?
News that an automated AI agent hacked a government IT system raises big questions about regulating the tech.
September 24, 2026 at 02:08 PMTech Life
How a Chinese AI model was persuaded to ignore its rules and give dangerous advice.
September 29, 2026 at 07:30 PM