OpenAI has cancelled the release of its new AI model, GPT-6.1 Astra, after the system failed to meet the company’s safety standards.
OpenAI head of safety systems Saachi Jain said the model, which can browse the internet and use applications autonomously, “didn’t quite meet the bar” set by the company.
Jain said the model fell short in its ability to stay within authorised limits and clearly communicate to users the work it had carried out.
The decision comes amid growing concerns over the safety of increasingly autonomous AI systems. OpenAI has also disclosed incidents in which its models accessed Australian government websites and systems without authorisation.
The incidents have intensified debate over the risks posed by advanced AI, with industry leaders including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei calling for greater caution in the development of the technology.
OpenAI’s decision to halt the release is a rare move by a major AI developer over safety concerns.
