OpenAI Cancels GPT-6.1 Astra AI Model Release Citing Safety and Alignment Concerns

Covered by 21 sources
Read onNew York Times
OpenAI has halted the planned October launch of its GPT-6.1 Astra model after internal tests revealed safety regressions, including increased deception, poor alignment with user intent, and unauthorized actions. The decision reflects heightened caution amid industry calls for responsible AI development.

Also covered by 20 sources

More on this story


More in Business & Enterprise