Stuart Russell gave Elon Musk’s legal team its clearest technical warning yet in the trial over OpenAI’s future, but his testimony stopped short of answering the central legal question.
Russell, a University of California, Berkeley computer science professor and longtime AI safety researcher, took the stand Monday as Musk’s only expert witness called to address AI technology directly.
His testimony gave jurors a serious framework for thinking about artificial general intelligence (AGI) competition. It did not, on its own, show that OpenAI breached a specific founding commitment.
What Russell’s testimony showed
Russell told jurors and Judge Yvonne Gonzalez Rogers that AI development carries risks, including cybersecurity threats, misalignment, and winner-take-all pressure around AGI. He ultimately said there was a “tension between the pursuit of AGI and safety,” according to TechCrunch.
That testimony helped Musk’s attorneys explain why the race toward more powerful AI systems can make it harder to maintain caution.
Russell was not there to review OpenAI’s corporate documents in detail or assess the company’s internal safety policies. His role was to describe the broader pressure around AGI: If companies or countries believe the first actor to reach AGI may gain a decisive advantage, slowing down can look risky, too.
Russell’s background gave that argument credibility. He signed the March 2023 open letter calling for a six-month pause in advanced AI development, and he has long argued for tighter oversight of frontier AI labs.
A December 2025 RAND report on AGI competition describes a similar dynamic. The report modeled AGI competition as a prisoner’s dilemma and found that incentives remain aligned toward acceleration until actors have coordination mechanisms grounded in a shared understanding of AGI’s global risks.
What Musk still needs to prove
Musk’s lawsuit argues that OpenAI abandoned its original nonprofit, safety-focused mission as it grew into one of the most powerful companies in AI. That claim now sits beside another fact: Musk is also running xAI, a direct AI competitor pursuing AGI.
KQED reported that Judge Gonzalez Rogers pointed to that tension from the bench, saying, “It is also ironic that your client, despite these risks, is creating a company in the exact space.” KQED also reported that OpenAI’s attorneys used Musk’s posts and messages to question his motives in bringing the case.
Those details do not settle whether OpenAI broke its founding commitments. They do make Musk’s argument harder to present as purely a safety dispute. His attorneys still need to connect the concerns Russell described to specific OpenAI decisions, documents, or governance choices that a court can treat as a breach.
The trial has already shown how difficult that connection may be. Russell made the AGI race sound dangerous, but OpenAI’s cross-examination narrowed the legal relevance of his testimony. Musk’s case needs more than a warning about industry incentives. It needs evidence that OpenAI’s conduct crossed a legal line.
For enterprise AI buyers, the trial raises a familiar governance question in a legal context: how much control should companies exercise over increasingly capable AI systems?
OpenAI’s broader product direction is already shaping how companies evaluate its 2026 roadmap, while competition among leading AI companies continues to put pressure on speed, safety, and oversight.
Also read: Our practical Grok prompts show how xAI’s chatbot is being positioned for everyday workplace research and planning.