Elon Musk clashes in OpenAI trial
- Elon Musk finished three days of testimony in Oakland on April 30, clashing repeatedly with OpenAI lawyer William Savitt over what he knew. - Under cross-examination, Musk said xAI had “partly” used OpenAI models to train Grok, while also saying he ignored for-profit “fine print.” - The case could reshape OpenAI’s nonprofit-controlled structure, its Microsoft ties, and how AI labs talk about model distillation.
This is a corporate-control trial disguised as an argument about AI ethics. Elon Musk is trying to convince a federal jury in Oakland that OpenAI broke its original nonprofit mission and turned itself into a giant commercial machine. OpenAI is trying to show something simpler — that Musk knew plenty, changed his mind, and is now suing a rival. On April 30, that fight got personal, with Musk sparring on the stand with OpenAI lawyer William Savitt and ending three days of testimony that ran more than seven hours. (usnews.com) ### What is the case actually about? Musk sued OpenAI, Sam Altman, Greg Brockman, and Microsoft in 2024. His core claim is that OpenAI was founded in 2015 as a nonprofit meant to build AI for humanity, then drifted into a profit-seeking structure that benefited private players instead. He is asking for huge da(usnews.com)rst and remedies later if he wins on the facts. (abcnews.com) ### Why did Thursday matter? Thursday was the end of Musk’s testimony and the sharpest stretch of cross-examination. Savitt kept pressing Musk on emails and documents that discussed making money, limiting openness, or moving toward a for-profit setup. Musk pushed back hard, saying the questions were misleading and that he was being cu(abcnews.com)errupting him. The judge, Yvonne Gonzalez Rogers, did tell Savitt to let him answer one question, but she did not buy Musk’s broader complaint that the questioning itself was improper. (usnews.com) ### What was OpenAI trying to prove? Basically, OpenAI wants the jury to see Musk as informed, not duped. Its lawyers pointed to messages showing early discussion of closed-source products and revenue models, which cuts against Musk’s claim that he thought the organization would stay (usnews.com)fit mission would remain intact. That is a risky answer, because it turns the case into a fight over whether informal understandings matter more than written records. (msn.com) ### Why was the Grok admission such a big deal? Because it handed OpenAI a very awkward fact. Under questioning, Musk said xAI had “partly” used OpenAI models to train Grok through distillation — basically using one model’s outputs to help teach another model. That matters because la(msn.com)trayed an altruistic mission while also admitting his own AI company borrowed from OpenAI’s systems. (techcrunch.com) ### Did the courtroom drift into AI-doom philosophy? Yes, briefly — because Musk keeps framing OpenAI’s origin story around AI safety. He testified that OpenAI existed partly as a response to what he saw as reckless attitudes at Google, including an old clash with Larry Page over whether superintelligent AI could endang(techcrunch.com)e jury is not being asked to decide who is morally right about existential risk. It is being asked to decide what promises were made, what structure was intended, and whether OpenAI broke legal duties. (usnews.com) ### Why does Microsoft matter here? Because Musk’s complaint is not just about vibes or founding ideals. It is also about who got the upside. He argues OpenAI’s most valuable technology ended up feeding a commercial partnership with Microsoft instead of a public-benefit nonprofit. OpenAI and Microsoft say the (usnews.com)other deals too. Still, Microsoft’s role makes the case bigger than a founder feud — it is about whether nonprofit control can survive giant capital needs. (abcnews.com) ### So what is the real stakes question? The real question is whether OpenAI’s hybrid structure looks like a genuine nonprofit-led system or a nonprofit wrapper around a giant business. If Musk wins, OpenAI could face pressure to rewrite governance, licensing, or fundraising plans. If OpenAI wins cleanly, Musk’s case starts to look le(abcnews.com)rcing unusually public answers about how frontier AI labs are financed, controlled, and copied. (abcnews.com) ### Bottom line The headline clash was Musk versus a lawyer. But the substance is bigger — who controls OpenAI, what its founding promises meant, and whether the AI industry’s own rules hold up when competition gets brutal. (usnews.com)
