Anthropic researcher warns AI 'could kill us all by the end of the decade'

Summary

CNBC's Kate Rooney reports on former Anthropic researcher Jacob Coxon's warning that the AI industry is racing toward self-improving superintelligence and that AI could threaten humanity by the end of the decade. The segment covers Anthropic researchers' extinction-risk estimates, alignment and recursive self-improvement concerns, and how these risks may affect Anthropic's IPO and AI/data center sentiment.

• Former Anthropic researcher Jacob Coxon warns the AI race for superintelligence is outpacing safety work.
• Coxon says AI builders privately fear the technology could kill all humans by the end of the decade.
• Current Anthropic researchers estimate a greater than 10% chance of AI-caused human extinction within the decade.
• AI alignment and recursive self-improvement are cited as making future systems harder to control.
• Anthropic is in the IPO process and has made AI safety a core part of its identity.
• Kate Rooney notes data center pushback and AI unpopularity are emerging risks.
• AI labs continue competing globally with OpenAI and Google on powerful models.