OpenAI has announced it will not release its latest AI model after flagging safety risks during in-house testing, industry\u2019s latest move to slow the rollout of the controversial frontier technology.<\/p>\n
The AI giant\u2019s announcement on Monday came as debate continues about the potential for AI to do catastrophic harm following a slew of incidents involving AI agents going rogue.<\/p>\n
Recommended Stories list of 4 itemsend of list<\/p>\n
Saachi Jain, OpenAI\u2019s head of safety systems, said GPT-6.1 Astra had failed to meet company standards for acting in accordance with human wishes during internal testing.<\/p>\n
\u201cFor anything regarding safety and alignment, there\u2019s a trade off,\u201d Jain said in a statement provided to Al Jazeera.<\/p>\n
\u201cYou really do need to find what\u2019s the right line between staying within scope, but also avoiding laziness in terms of how the model actually pursues tasks even when it hits friction.\u201d<\/p>\n
While GPT-6.1 Astra improved from its predecessor in some areas, Jain said, the model did not meet the bar for \u201cscope and authorization, and how it communicates back to the user about the type of work it\u2019s done\u201d.<\/p>\n
\u201cOf course we want to make sure our model development is safe no matter whether that\u2019s in the company, or when we ship it to users,\u201d Jain said.<\/p>\n
\u201cBut when we ship it to users, we have an extremely high bar in terms of safety and alignment.\u201d<\/p>\n
The decision, announced on the eve of OpenAI\u2019s annual developer conference in San Francisco, was first reported by The Wall Street Journal.<\/p>\n
Fears of AI escaping human control have prompted industry-wide calls for a slowdown in development to allow researchers time to implement stronger safeguards.<\/p>\n
In an influential essay earlier this month, Dario Amodei, the CEO of Claude creator Anthropic, called on AI developers to \u201cpace the frontier\u201d to mitigate the risk of catastrophic harm.<\/p>\n
While Amodei\u2019s call received the backing of rivals, including OpenAI CEO Sam Altman and xAI chief Elon Musk, other key industry figures, such as Meta boss Mark Zuckerberg, have dismissed the need for a coordinated slowdown.<\/p>\n
The risk of AI models going rogue has been in the spotlight since July, when OpenAI revealed that its models had broken out of a controlled testing environment and hacked the software start-up Hugging Face.<\/p>\n