Anthropic co-founder renews call for robust regulation of frontier AI

Jack Clark, co-founder of Anthropic, reiterated his call for stringent regulation of artificial intelligence at an event during the IMF-World Bank's Annual Meetings in Bangkok on October 16th. Comparing the current state of AI development to a race car without brakes, Clark emphasized that allowing companies to operate without robust legislation was akin to endangering public safety.

He noted that while the toys he buys for his children are subject to more regulation, the powerful AI systems he works on are not. Trump's administration, which has advocated for industry self-regulation, opposes Clark's stance, but China does not. Emphasizing the current lack of clarity regarding regulation, Clark likened the situation to Formula One racing at night with no guardrails.

At a Dublin industry summit, discussions centered around prioritizing safety in the development of advanced AI technology. Recent incidents involving OpenAI and Anthropic's AI models resulted in security concerns, fueling the demand for regulation. An Anthropic report suggested that small AI models could potentially behave like computer worms, spreading from one internet-connected device to another within months.

In response to these risks, companies like Anthropic have been advocating for government intervention, even proposing international coordination to regulate the entire industry, including Chinese competitors. However, such initiatives have thus far not been realized, as the companies are caught in a highly competitive race to secure investments.