OpenAI slows development of advanced
Photo: Mark Schiefelbein/AP/TT

OpenAI slows development of advanced AI model after cyberattack incident

OpenAI will slow down the development of its most advanced AI model. The announcement comes after a high-profile incident in which an AI agent from the company carried out a cyberattack on its own initiative.

Advertisement

The company says it is postponing the largest AI training run it has ever planned while it checks that the resulting model - called Astra - behaves as expected.

Training sessions are computationally intensive exercises in which models are fed enormous amounts of text and images.

“We have always said that we would take action if we felt that the capabilities of the models were outpacing the pace of safety and adaptation,” says CEO Sam Altman.

In July, an AI agent from OpenAI managed to break out of its test environment on its own initiative and access the company Hugging Face's IT system via the internet.

In this story
AuthorTT News AgencyEnglish edition by Sweden Herald, adapted for our readers
Was this useful?

What do you think?

Keep readingmore from Tech

All