Altman Draws Regulatory Line Between OpenAI and Anthropic Regarding Risk

Sam Altman, CEO of OpenAI, said that society must accept a degree of risk to reap the benefits of artificial intelligence, describing his company's position as leaning toward lighter regulation than the proposals put forward by Anthropic. This came during an interview on POLITICO's Decoded podcast.

Altman added that he would not accept any deal promising an environment completely free of serious breaches, misuse, or fraud, arguing that the beneficial uses of AI will far outweigh its harmful uses. At the same time, he emphasized that this does not mean accepting catastrophic risks, including the loss of human control over AI.

Altman pointed to criticisms directed at the proposals of Anthropic CEO Dario Amodei, noting that they could lead to something resembling regulatory capture. Anthropic responded that its proposed rules would only apply to frontier models, which are the most advanced and capable.

This divergence between the two companies comes as global debates intensify over how to regulate the most advanced models. The dispute reflects a difference in approach between calls for stricter legislation and others favoring a flexible framework that fosters innovation.

What do these terms mean?

Light-touch Regulation: An approach that favors limited, flexible rules over strict constraints, aimed at not hindering innovation. Regulatory Capture: A situation in which major companies control the drafting of industry regulations to serve their own interests. Frontier Models: The most advanced and capable AI models, which represent higher risks than others.