An AI researcher who has worked at both Anthropic and Open AI has claimed that AI staff are “genuinely frightened” about the progress of AI and the potential danger that might pose for humanity according to the BBC.
The British AI researcher Jacob Coxon recently achieved a level of fame and notoriety when he announced via X ,formerly Twitter, his departure from Anthropic over his concerns that both Anthropic and OpenAI were ”gambling with our lives” over their rush to advance their AI models.
He claimed in one of his companion posts that ”the people building AI earnestly believe that it could kill us all by the end of the decade”. His initial post received over 100 million views in just 18 hours and nearly 170 million at the time of publishing of this article.
The subsequent virality of his post and news media coverage has brought a wave of attention to the danger of AI which in turn also caused a shockwave across the AI industry with renewed calls for regulation or a slower pace of development.
These calls are themselves not new but with the subsequent unprecedented spike in focus on AI concerns, major stakeholders are now attempting to capitalize on this attention while people globally are paying more attention than they ever have.
In an essay titled We Must Pace the Frontier, published on Saturday Dario Amodei ,the head of Anthropic, said he is concerned that since this summer, AI has been “advancing drastically faster”.
The Independent UK reported that Amodei expressed concern that with the current growth of AI it is possible that within 6 – 12 months that we could see an AI swarm that takes over the entire internet. This follows several incidents of hacking with the first major reported hacking incident being in July involving Open AI and another AI developer Hugging Face.
In his essay the Anthropic CEO shared a three part plan on how to effectively and responsibly curb the rate of development of AI.
“I believe that if slowing down bought us even an extra year or two before models reach critical levels of capability, and we used that time to advance alignment, we could greatly reduce the risk that something goes seriously wrong,” Mr Amodei said.
In response to Amodei’s post on X, Elon Musk, owner of X, and SpaceXAI said simply ”Dario is right”. He was also joined by another industry head Sam Altman.
Open AI chief executive Mr Altman wrote on the social media site: “I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we’ve had at OpenAI in recent weeks.”
Their support comes in the face of a call from Amodei for independent evaluators to have employee level access to report issues and monitor the safety practices of AI developers and researchers.
Coxon whose mega-viral post seemed to spawn this recent flurry of news articles and social media responses said in a BBC interview that while he clearly supports a slowdown it would need to be coordinated with China to avoid a ”race at international scale”.
