AI Generated Image
Anthropic CEO Dario Amodei: "For too long the industry lied" about AI risks
CBS News • • general
Key Points:
- Anthropic CEO Dario Amodei acknowledges significant risks in AI development, emphasizing the need for the tech industry to be honest about these dangers and to slow the pace of AI advancement.
- Former Anthropic researcher Jacob Coxon warns that self-improving superintelligence could pose an existential threat to humanity, likening the risk to scenarios depicted in science fiction like "Terminator."
- Amodei proposes rigorous testing of AI models before release and plans to grant independent evaluators permanent access to verify safety practices, likening this oversight to food inspection.
- While supporting federal regulation and measures like an AI "kill switch," Amodei opposes an outright ban on superintelligence, citing potential medical and societal benefits and concerns over international competition, especially with authoritarian regimes.
- Amodei expresses discomfort that AI is being developed by private companies rather than governments, advocating for joint governance by democratically-elected governments to ensure responsible oversight and reduce risks.
