Elon Musk is testifying in his legal battle against OpenAI, claiming the company breached a charitable trust by converting from a nonprofit into a for-profit enterprise. The core argument is straightforward. OpenAI was founded as a nonprofit for the benefit of humanity. It used that status to recruit talent and secure tax exemptions. Then it changed course. Musk wants the restructuring unwound.
Gary Marcus, the NYU Professor Emeritus who has been following the trial closely on his Substack, Marcus on AI, believes Musk has a legitimate point but may be his own worst enemy in making it. If the trial becomes a referendum on Musk rather than on what happened at OpenAI, he could lose.
The nonprofit question
The legal questions are complex, but the underlying issue is clear. OpenAI presented itself as a mission-driven organisation, attracted people on that basis, and then shifted to a structure designed to generate returns for investors. Marcus argued this amounts to exploitation of the people who worked for the nonprofit and of the public trust that enabled it.
OpenAI's original promises to California and to the broader research community created obligations. If the company were still operating as a nonprofit, Marcus suggested, it would be under pressure to wind down products with harmful consequences, such as tools used to generate non-consensual deepfake material, and to prioritise work that benefits humanity rather than shareholders.
Musk keeps getting in his own way
The problem is execution. OpenAI's lawyers have proved skilled at getting under Musk's skin during cross-examination. Rather than keeping the focus on OpenAI's early history and the circumstances of its conversion, Musk has made sweeping declarations about AI safety without being able to answer basic follow-up questions. Marcus described this as embarrassing.
The goal, Marcus argued, should not be a personal victory for Musk. It should be a legal remedy that voids the nonprofit-to-profit transition. But Musk's ego keeps pulling the trial back toward himself, which is exactly where OpenAI's defence team wants it.
Altman's turn on the stand
Sam Altman is expected to testify, and Marcus was blunt in his assessment. He described Altman as someone who projects insincerity and has a track record of making statements that later prove false, including claims about his equity in OpenAI and his position on compensating artists whose work was used to train the company's models.
Related reading
- Musk drops 24 of 26 claims against OpenAI three days before trial, gambling everything on two
- X rolls out Grok-powered translation and AI image editing tools
- Altman vs Sutskever II: Do We Now Know What Ilya Saw?
OpenAI's defence is likely to argue that any deals Altman struck with companies doing business with OpenAI were in the organisation's best interests and involved no conflict of interest. It will fall to Musk's legal team to use Altman's own past statements and deposition testimony to challenge that account.
The complication running through the entire case is that Musk owns xAI, a direct competitor to OpenAI. Even if he gains nothing directly from the nonprofit remedy, he benefits indirectly from anything that constrains his rival. Both sides have valid arguments. Neither protagonist inspires much public trust.
