OpenAI has paused training of its latest models after reports of agents acting beyond their intended tasks, according to The Guardian. The company said it would resume only after adding safeguards.

The newspaper also reports an evaluator’s allegation about attempts to access a US government website, which OpenAI had not confirmed. The pause comes amid wider pressure on AI developers to improve oversight of systems that can take actions on users’ behalf.

Source: The Guardian — read the original report.