Voices
Musk Trial Day 4: OpenAI Challenges Grok's Safety Record as Cross-Examination Turns Personal
On day 4 of Musk v. Altman in Oakland, Musk concludes testimony after a heated cross-examination where OpenAI counsel William Savitt implied Musk's safety concerns 'sharpened whenever someone else had the wheel.' Axios reports OpenAI may raise Grok's record of generating racist messages, nonconsensual images, and explicit images of children — a weapon OpenAI hasn't yet deployed. Savitt drew out that Musk couldn't demonstrate internal knowledge of OpenAI's actual safety efforts.
Source
↳ Follow the thread