OpenAI trial expert witness warns of AGI arms race risks

In the US legal battle between Elon Musk and OpenAI, the plaintiff’s legal team is focusing on the core purpose of the artificial intelligence organization. Musk’s attorneys argue that the organization was set up as a charity focused on AI safety and lost its way in pursuit of lucre. Today, the plaintiff called Stuart Russell, a computer science professor at the University of California, Berkeley, to testify as an expert witness. Russell’s role was to establish that this technology is dangerous enough to worry about. He testified that there are risks associated with the development of AI, ranging from cybersecurity threats to problems with misalignment and the winner-take-all nature of developing artificial general intelligence (AGI). This tension is central to the trial, as virtually every one of the OpenAI founders have warned about AI risks while simultaneously attempting to build AI as fast as possible.

Russell has long been a critic of the global arms-race dynamic among artificial intelligence labs. In March 2023, Russell signed an open letter calling for a six-month pause in AI research—a document that Musk also signed. Although Russell’s broader concerns regarding the existential threats of unconstrained AI were limited in open court by Judge Yvonne Gonzalez Rogers following objections from OpenAI’s attorneys, his testimony highlighted the friction between rapid development and safety.

The debate over these warnings extends beyond the courtroom into US policy. For instance, US Senator Bernie Sanders has pushed for a law imposing a moratorium on data center construction, drawing on similar AI safety concerns. However, these warnings from industry figures have met with skepticism. Hodan Omaar, who works at the trade organization the Center for Data Innovation, questioned the selective use of these warnings. Omaar noted that “it is unclear why the public should discount everything tech billionaires say except when their words can be recruited to fill gaps in a precarious argument.”

Why it matters

The trial exposes the fundamental conflict between the original safety-focused mission of AI labs and the massive compute spend—the expenditure on computing resources required for AI training—needed to build artificial general intelligence (AGI). This capital requirement forces a pivot to for-profit models to secure investment.