Concerns About AI’s Potential Threat Highlight Need for Regulation

Jacob Coxon, a former researcher at Anthropic, discussed his concerns about artificial intelligence (AI) during a recent appearance on ‘Special Report’ with Fox News. He emphasized the potential risks AI poses, highlighting a fear that without regulation, AI might endanger humanity in the next decade.

Coxon shared his resignation statement on X, expressing that tech companies are irresponsibly advancing towards self-improving AI. His main worry is that such advancements gamble with lives. Having worked at OpenAI and Anthropic for three years, Coxon believes regulation could prevent this scenario.

During an interview with Bret Baier, Coxon stated that the threat of AI is manageable. He claimed there’s more than a 10% chance that AI could jeopardize humanity, but with careful action, the risk can be reduced. He urged international cooperation to ensure AI does not lead to extinction. Coxon noted a global call for regulation, stressing that the problem is solvable with concerted effort.

Evan Hubinger, who leads alignment science at Anthropic, supported Coxon’s sentiments on X. Hubinger concurred that AI poses a real threat and expressed confidence that it’s above 10% likely within the next decade. He noted that although current AI models pose a low risk, superintelligence through self-improvement remains a concern. According to Hubinger, the process is developing quicker than anticipated.

Coxon agrees with Hubinger that the immediate danger is low, but he warns that recursive self-improvement could rapidly escalate, possibly beginning as early as next year. This apprehension is shared across the industry, where many worry their work might accelerate AI takeover by decade’s end.

AI’s current trajectory sees it utilized internationally as a ‘superweapon’ in an arms race. Coxon describes the technology as potentially the most dangerous humanity has devised. He calls for worldwide collaboration to avoid catastrophic outcomes.

Although many tech leaders acknowledge the need for moderation in AI development, their calls are often not heeded. Coxon emphasizes that the call for regulation is urgent and genuine, not merely a marketing tactic.

Despite the risks, Coxon acknowledges AI’s potential for substantial discoveries. Yet, he urges for a careful approach to avoid hastened developments. Addressing Baier’s question about solving the issues, Coxon remained optimistic, suggesting that with deliberate effort, a solution is achievable.

Fox News Digital sought Anthropic’s comments on these concerns. For further feedback or tips, contact Sydney Shea, editor at Fox News Digital.