World · Daily Terminal
Andrey Rudakov | Bloomberg | Getty Images There is more than a 10% chance that artificial intelligence could 'kill all humans,' an Anthropic safety researcher said on Tuesday, hours after another employee resigned over concerns that AI labs are 'gambling with our lives.' Jacob Coxon, a researcher at Anthropic, stated that neither Anthropic nor OpenAI is acting responsibly. 'They are racing straight to self-improving superintelligence and gambling with our lives,' Coxon said in a post on X. Self-improvement means AI systems can improve themselves without much human intervention. Recursive self-improvement, though not yet possible, is the goal AI labs are working toward.
Coxon warned, 'Do not underestimate the power of this technology. These will soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.' He added that people building AI earnestly believe it could kill us all by the end of the decade. Evan Hubinger, an alignment science lead at Anthropic, agreed, stating that Coxon's statement was correct and that Anthropic has no plan for this scenario. 'I personally think it is >10% within the next decade,' Hubinger said.
In June, Anthropic acknowledged that 'full recursive self-improvement also might increase the risks of humans losing control over AI systems.' Concerns about out-of-control AI are not new; Elon Musk has repeatedly warned about this threat. Major researchers and academics have also expressed alarm over companies losing control of AI systems. A recent incident where an OpenAI model breached Hugging Face, a major platform for open-source developers, has heightened these concerns.
Coxon cited this incident as a 'warning shot,' suggesting that U.S. labs are more likely to coordinate. However, he warned that a global AI race is inevitable and might require costly actions like a temporary ban on improving model capabilities.
Source: CNBC
