The most revealing moment of Musk v. Altman's first week wasn't the doomsday warning - it was the admission about training data.
Elon Musk took the stand in a black suit and tie and argued that OpenAI CEO Sam Altman and president Greg Brockman had deceived him into bankrolling what he believed would be an open, nonprofit AI research lab. He warned that AI could kill humanity. Then he acknowledged that xAI - his own AI company and the maker of the Grok chatbot - had used distillation from OpenAI's models during training.
Distillation means using one model's outputs to train another: feed your system the responses that GPT-4 generates, and use those as training data to make your own model stronger. OpenAI's terms of service explicitly prohibit using their outputs to train competing AI products. Musk is suing OpenAI for betraying its founding principles while leading a company that, by his own account, built on OpenAI's proprietary outputs to get started.
The lawsuit's central argument is that Altman and Brockman misrepresented OpenAI's future direction to secure Musk's early donations, which he frames as a commitment made under false premises. ChatGPT, which launched in late 2022 and turned OpenAI into one of the most valuable private companies in the world, sits at the center of what Musk says went wrong.
OpenAI has released internal communications it says show Musk endorsed the commercial structure he now opposes in court. The company also filed counterclaims of its own. The distillation admission has no clean resolution for Musk: even if OpenAI doesn't pursue it as a formal counterclaim, it sits on the record as evidence that his company benefited directly from the same technology he claims was mishandled.