Anthropic Researcher Warns About AI Dangers

Jacob Coxon, a former researcher at Anthropic, recently voiced concerns about the potential dangers of artificial intelligence (AI). He fears that the relentless pursuit of self-improving AI might gamble with human life. Speaking on ‘Special Report’ with Fox News, Coxon emphasized the urgency of implementing regulations to prevent catastrophic outcomes.

Aged 27, Coxon has experience in pre-training research at both OpenAI and Anthropic. He announced his resignation on X, highlighting that tech companies are not acting responsibly, thus risking society’s future.

Coxon stated, ‘It’s not a hopeless situation. The problem is eminently solvable.’

During his interview with Fox News chief political anchor Bret Baier, Coxon explained the need for immediate action to mitigate a greater than 10% chance of AI threatening humanity within the next decade. He stressed the importance of international cooperation to prevent racing into potential extinction without necessary precautions.

Evan Hubinger, alignment science lead at Anthropic, supported Coxon’s concerns. He mentioned on X that superintelligence from recursive self-improvement might pose a serious threat in the near future. Hubinger reiterated that while current AI models pose low risk, the rapid development of superintelligence warrants careful attention.

Coxon described today’s technology as a new “superweapon” in a global arms race, urging the need for global oversight. He noted that sector leaders have called for moderation and a slow-down on AI development, an appeal that needs to be taken seriously.

Despite the looming threats, Coxon acknowledged the potential benefits AI could bring if developed cautiously. He firmly believes that with proactive measures, the challenges associated with AI can be addressed, avoiding a disastrous path.