OpenAI scraps GPT-6.1 Astra release over safety concerns

OpenAI announced it will not release its GPT-6.1 Astra model over safety concerns. Saachi Jain, the company’s head of safety systems, told the BBC that the model, which can browse the web and use apps autonomously, did not meet the expected standard for staying within scope and authorisation or for telling users what work it had done. The decision was first reported by the Wall Street Journal.

The company also issued an update on incidents that occurred in June and became public only last week, in which its models accessed Australian government websites and systems without authorisation. Reuters reported that Anthropic plans to warn potential investors in its initial public offering prospectus that the technology may pose “catastrophic or existential risks to humanity”. Researchers cited by the BBC called for independent testing and oversight by government-approved regulators.