Elon Musk testifies in OpenAI trial
Business Insider reports he cites Obama meetings and AI-as-child analogy, trust is argued in court while model control sits in contracts
Images
Elon Musk took the witness stand this week as the first witness in a California trial pitting him against OpenAI chief executive Sam Altman, according to Business Insider. In testimony described by the outlet, Musk argued he has been consistent about artificial intelligence risk for years, comparing AI to “a very smart child” that needs values instilled early. He cited past conversations with Google co-founder Larry Page and meetings with former US president Barack Obama as proof that he raised alarms long before today’s boom.
The dispute lands in the middle of an industry where personal histories are routinely repackaged as governance credentials. AI companies now sell “safety” as a product feature to regulators, enterprise customers, and cloud partners, while simultaneously racing to lock up the practical inputs that matter: compute capacity, datacentre space, distribution, and long-term contracts. Courtroom narratives about who warned whom can read like a proxy fight over who gets to define “responsible AI” while still controlling the pipes.
Musk’s appearance also highlights how quickly alliances in the sector have hardened into legal positions. The firms building large models depend on suppliers and platforms that can raise prices, ration capacity, or attach conditions; the suppliers want guarantees that justify multibillion-dollar infrastructure bets. When those relationships sour, yesterday’s co-founder story becomes tomorrow’s exhibit list.
The case also shows why the public often learns about AI governance through litigation rather than disclosure. The most valuable details—internal communications, board deliberations, and the terms that govern access to models and computing—tend to surface only when parties are compelled to produce documents. Outside court, much of the industry’s risk management is communicated through press releases and carefully scoped “principles,” while the binding obligations sit in contracts.
Business Insider’s account of Musk’s testimony focused on his attempt to rebut claims about his relationship with Altman and to frame his long-standing posture on AI. The trial continues, but the opening image is already set: one of the sector’s most influential entrepreneurs making the case for trustworthiness not with audits or enforceable constraints, but with recollections of private meetings.
Musk told the court he warned Obama about AI. The companies now building it are spending far more on compute and lawyers than on proving which warnings changed anything.