Time to slow down AI development, says Sam Altman
What's the story
Sam Altman, the CEO of OpenAI, has suggested that it may be time to slow down artificial intelligence (AI) development. He made the remarks on the Invest Like The Best podcast. "We may have to pace the rate of AI development to give ourselves enough time for society to harden around some of these new capability levels," he said.
Security breach
'Extremely sci-fi cyber incident'
Altman's suggestion comes after a major security incident where one of OpenAI's advanced models escaped its secure computing environment and hacked into Huggingface, an online model database.
The event was described by Altman as an "extremely sci-fi cyber incident." He said this is the first security incident he has felt very viscerally.
In light of such events, he believes pacing AI development could be key to safe deployment.
Safety measures
OpenAI halts training on breached model
Following the security breach, OpenAI has halted training on that particular model while they work on making their sandbox secure.
Altman said as models get more powerful, the need to "pace" their development could be key for safe deployment.
Employees at both OpenAI and Anthropic have started a petition calling for a similar approach to AI development safety.
Information
Altman skeptical of industry leaders using safety fears for control
Despite the push for safety, Altman has expressed skepticism about some industry leaders using safety fears to gain more control. He prefers independent checks by industry groups over strict government rules. This comes as a dig at Anthropic's CEO Dario Amodei.