Musk sues OpenAI, accuses Altman of betraying its mission

Elon Musk has sued OpenAI, its CEO Sam Altman and its chairman Greg Brockman. The organization’s co-founder claims the company has breached the agreement that gave rise to OpenAI: developing advanced artificial intelligence for the benefit of humanity, rather than to maximize revenue.

The lawsuit, filed in a San Francisco court, turns a dispute over OpenAI’s direction into a legal matter. It also puts the company’s relationship with Microsoft, its main technology and financial partner, at the center of the case, along with the growing secrecy surrounding its most powerful models.

From open lab to capped-profit company

OpenAI was founded in December 2015 as a nonprofit organization. Musk helped create it alongside Altman, Brockman, Ilya Sutskever, Wojciech Zaremba and other researchers and entrepreneurs. Its stated purpose was to conduct AI research in a way that benefited humanity as a whole.

At that stage, the name was no small matter. OpenAI published research, code and models, in contrast to the more closed strategy pursued by major labs such as Google and Facebook. The idea was that a technology capable of transforming the economy, science and security should not remain under the exclusive control of a single company.

In 2019, OpenAI created a capped-profit subsidiary, OpenAI LP, to attract the capital needed to train large-scale models. The nonprofit entity, OpenAI Inc., formally remained in control of the structure. The company argued at the time that developing advanced systems required investments of billions of dollars in chips, data centers and specialized staff.

That change is at the heart of the dispute. Musk claims that the capped-profit model and close collaboration with Microsoft have displaced the organization’s founding goal. His allegations will have to be proven during the court proceedings.

GPT-4 and Microsoft at the heart of the allegations

The lawsuit presents GPT-4 as evidence of that shift. OpenAI launched the model in March 2023 and made it available to paying customers through ChatGPT Plus and its developer interface. Unlike earlier models from the organization, OpenAI did not publish GPT-4’s essential technical details, such as its architecture, training data or size.

Musk argues that OpenAI has treated GPT-4 as proprietary technology, contradicting its promise to work openly. He also alleges that the model is approaching artificial general intelligence, or AGI: AI capable of solving very different intellectual tasks at a human-comparable level. There is no technical definition or universally accepted test for determining when a system reaches that category.

Microsoft is relevant for two reasons. The company announced a $1 billion investment in OpenAI in 2019 and expanded the relationship in 2023 with a new multiyear, multibillion-dollar investment. Azure, Microsoft’s cloud-computing platform, also provides the infrastructure OpenAI uses to train and serve its models.

The deal has helped bring ChatGPT and GPT-4 to hundreds of millions of users and integrated OpenAI technology into products such as Bing and Microsoft 365. But it has also fueled debate over whether OpenAI retains real autonomy from one of the world’s largest technology groups. Microsoft is not named as a defendant in this case.

What Musk is asking the court to do

Musk is seeking an order requiring OpenAI, Altman and Brockman to stop benefiting from technology developed by the organization when that technology falls outside its original nonprofit purpose. He also asks the court to bar OpenAI and Microsoft from using models such as GPT-4 for commercial purposes if they are found to have reached the level of AGI contemplated in the founding agreements.

The lawsuit includes claims for breach of contract, breach of fiduciary duty and unfair competition. The initial question will be whether the documents and conversations between the founders constitute a contract enforceable in court and what specific obligations they imposed on OpenAI.

Musk left OpenAI’s board in 2018. Since then, he has repeatedly criticized the organization and, last year, launched xAI, a rival company developing the Grok assistant. That rivalry does not by itself invalidate his legal arguments, but it adds clear business context to a lawsuit involving one of the industry’s most influential labs.

A debate that goes beyond OpenAI

The case pits two visions of AI development against each other. One prioritizes open research and broad distribution of its capabilities. The other holds that the most advanced models require such vast resources and pose such serious risks that they need corporate structures, restricted access and safety controls.

OpenAI will have to respond to the allegations in court. In the meantime, the lawsuit leaves the entire industry with an uncomfortable question: when an organization created to conduct research for the public good needs billions of dollars to compete, what mechanisms ensure that its mission remains a priority over its commercial interests?

The case may take years to resolve, but the proceedings already require scrutiny of a tension that has become central to the AI race: who funds the most capable models, who controls them and under what rules they are made available to the public.