OpenAI scraps release of latest AI model over safety concerns | Technology News

DEVELOPING STORYDEVELOPING STORY,

AI giant says GPT-6.1 Astra failed to meet alignment standards during internal testing.

OpenAI has cancelled the release of its latest AI model over safety concerns, the latest move by industry to slow the development of the frontier technology.

The AI giant’s announcement on Monday comes amid heightened fears about the potential for AI to do catastrophic harm following a slew of incidents involving AI agents going rogue.

Saachi Jain, OpenAI’s head of safety systems, said the upcoming model, GPT-6.1 Astra, had failed to meet its standards for acting in accordance with human wishes during testing.

“For anything regarding safety and alignment, there’s a trade off. You really do need to find what’s 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,” Jain said in a statement provided to Al Jazeera.

While GPT-6.1 Astra improved in comparison to its predecessor in some areas, the model did not meet the bar for “scope and authorization, and how it communicates back to the user about the type of work it’s done,” Jain said.

“Of course we want to make sure our model development is safe no matter whether that’s in the company, or when we ship it to users,” Jain said.

“But when we ship it to users, we have an extremely high bar in terms of safety and alignment.”

More to follow…