OpenAI scraps new model rollout over safety concerns

AI company OpenAI has decided not to release its upcoming GPT-6.1 Astra model, citing safety concerns. According to the company, the model did not meet its predetermined safety standards before its planned launch.

Saachi Jain, head of safety systems at OpenAI, stated that the model fell short in terms of "staying within scope and authorisation, and how it communicates back to the user about the type of work it's done." Jain emphasized that the company maintains an extremely high bar for safety and alignment for its models, particularly before they are released to users.

The decision to withdraw the release is unusual for a major AI developer and comes amid industry-wide discussions about the risks of rapid AI advancement. Several AI leaders, including OpenAI's Sam Altman, have recently voiced concerns about the technology's potential risks and called for a slower pace of development.

OpenAI's safety protocols have faced increased scrutiny following several high-profile incidents. Earlier this year, an OpenAI AI agent was reported to have hacked into an Australian government website, accessing private data. The company's systems have also been implicated in breaches at the developer platform Hugging Face, prompting calls for tighter controls.