When it comes to legal disputes, Elon Musk’s definition of victory may not always be winning in court.
Last week, Musk sued OpenAI and co-founders Sam Altman and Greg Brockman for breach of contract and fiduciary duty. Experts say the case is built on a questionable legal foundation, because the contract at the heart of the suit isn’t a formal written agreement that was signed by all parties involved.
Rather, Musk is alleging that the early OpenAI team had set out to develop artificial general intelligence, or AGI, “for the benefit of humanity,” but that the project has been transformed into a for-profit entity that’s largely controlled by principal shareholder Microsoft.
Musk used much of the 35-page complaint (plus attached exhibits) on Friday to tell his side of the story and to remind the world of his central position in the creation of a company that’s since become one of the hottest startups on the planet, (OpenAI ranked first on CNBC’s Disruptor 50 list in 2023) thanks largely to the viral spread of ChatGPT.
“It’s certainly a good advertisement for the benefit of Elon Musk,” Kevin O’Brien, partner at Ford O’Brien Landy LLP and former assistant U.S. attorney, told CNBC. “I’m not sure about the legal part though.”
O’Brien, who isn’t involved in any cases with Musk, added, “One thing that jumped right out at me is there’s no contract.”
In the suit, Musk’s lawyers say they want OpenAI to return to its work as a research lab and no longer exist for the “financial benefit” of Microsoft. Musk, who’s worth over $200 billion, is unconcerned about the legal costs of floating a suit that has no clear personal economic benefit and is of questionable merit.
Shannon Capone Kirk, global head of e-discovery and AI for Ropes & Gray LLP, told CNBC that Musk might just be seeking to force more information into the public realm about how OpenAI has been operating and how its business objectives have morphed in recent years.
It’s a “high-profile case with great public interest, a consequence of which might lead to OpenAI being available to everyone,” said Kirk, who isn’t working on any cases involving Musk. “Is that the real objective?”
In their complaint, Musk’s attorneys allege that OpenAI “has been transformed into a closed-source de facto subsidiary of the largest technology company in the world: Microsoft.” They also say the arrangement goes against a founding agreement and 2015 certification of incorporation that OpenAI established with Musk, who was a pivotal donor to OpenAI in its early years.
Musk’s attorneys said their client contributed over $15 million to OpenAI in 2016, which was “more than any other donor” and helped the startup build a team of “top talent.” The next year, Musk gave nearly $20 million to OpenAI, which the attorneys reiterated was more than other backers. In total, Musk invested over $44 million into OpenAI from 2016 through September 2020, according to the suit.
The lawsuit fits a pattern for Musk, who has frequently posted on X and commented in public forums about his importance to the creation of OpenAI.
In November, Musk told an audience at the The New York Times’ DealBook conference that OpenAI had deviated from its original mission.
“OpenAI should be renamed ‘super closed source for maximum profit AI,’ because this is what it actually is,” Musk said onstage at the event. He noted that it’s transformed from an “open source foundation” to multibillion-dollar “for-profit corporation with closed source.”
Is there injury?
In the suit, Musk’s attorneys allege that the inner workings of OpenAI’s GPT-4 AI model are “a complete secret except to OpenAI—and, on information and belief, Microsoft,” and that the secrecy is driven by commercial gain rather than safety. Musk has publicly beefed with Microsoft for a while, and in May 2023, Musk attorneys accused the company of using X (formerly Twitter) data in unauthorized ways.
Even if OpenAI’s mission has changed, that doesn’t mean Musk has a solid legal case.
“If he has any hopes to recover, he’s going to have to prove that there was this agreement – that the company be open and not for profit and all these other things, and that the failure to do so has caused him injury, which is a separate problem,” O’Brien said. “It’s hard to see where the injury is here.”
Musk’s attorneys didn’t respond to a request for comment.
Musk has an AI company of his own, X.AI, which introduced a competing chatbot called Grok in November after two months of training. In December, X.AI filed with the SEC to raise up to $1 billion in an equity offering. And Musk is also developing autonomous vehicle tech and humanoid robotics, which require AI advances, at Tesla.
He’s been known to hire bigwigs from OpenAI, poaching Andrej Karpathy, a former OpenAI software engineer, over to Tesla in 2017. More recently, Musk hired Kyle Kosic from OpenAI to join X.AI.
One of Musk’s goals with this case, lawyers said, may be to shed light on details of OpenAI’s GPT-4 in the discovery process, should it get that far. O’Brien said it can be tough to keep intellectual property and other internal details private when a lawsuit is brought.
Kirk agreed, saying that in the discovery stage, there may be “lots of document requests for all kinds of communication,” such as internal conversations, text messages and more. Some of the documents produced may come with protective orders that keep them out of the public.
A portion of Musk’s lawsuit rests on the idea that OpenAI has already reached AGI, typically defined as AI that can operate on the same level — or higher — than humans when completing a wide array of cognitive tasks. The suit claims that since GPT-4 is “better at reasoning than average humans” based on test scores on the Uniform Bar Exam, GRE Verbal Assessment and even the Advanced Sommelier exam.
As part of its contract with OpenAI, Microsoft only has rights to OpenAI’s “pre-AGI” technology, and it’s up to OpenAI’s board to determine whether the company has reached that milestone.
In a memo to employees on Friday following the lawsuit, OpenAI said that “GPT-4 is not an AGI.”
“Importantly, an AGI will be a highly autonomous system capable enough to devise novel solutions to longstanding challenges,” Chief Strategy Officer Jason Kwon wrote. “GPT-4 can’t do that.”
Much of the AI community is in agreement with Kwon.
Kirk said “part of what they’re going to be litigating” is the question of what is AGI.
Read the full complaint here: