Top AI researchers call for regulation of self-improving AI

Per TechFlow, on September 28, according to The Wall Street Journal, the heads of research at OpenAI, Anthropic, Meta Platforms, and Microsoft are calling on policymakers to assess the extent of their respective companies' progress in automating AI research, in response to industry-wide calls for stricter oversight of AI technology.

In a paper released on Monday, researchers including OpenAI Chief Scientist Jakub Pachocki, Anthropic Co-founder Jack Clark, Microsoft Chief Scientific Officer Eric Horvitz, and Meta Vice President of Artificial Intelligence Research Dawn Song wrote that the automation of AI research could trigger an "intelligence explosion," a development that could compress progress that would normally take years into just months or even less time, thereby causing technological evolution to outpace human comprehension. Writing in their personal capacities, the paper's authors noted that policymakers should urgently demand greater transparency to understand how far each company has progressed in automating model development.