Elon Musk’s chances of winning his second lawsuit against OpenAI look grim, but in law nothing is entirely hopeless.
The entrepreneur has claimed the nonprofit he helped found cannot legally convert to a corporation without violating the very purpose for which it was created nine years ago—to benefit humanity as a whole by developing the world’s first artificial general intelligence (AGI).
Musk is seeking to force his former creation to compensate him for treble the $44.6 million he donated over five years and compel it to open-source all the research findings behind its neural network GPT-4.
The latter coincidentally would also serve the interests of his own xAI research lab, a direct competitor to OpenAI.
Yet U.S. law isn’t kind to private litigants like Musk seeking redress against charitable nonprofits to which they have made tax-deductible donations. Attempts made after the fact to demand money back or insist funds be used differently are typically doomed.
“All those cases fail,” Brian Quinn, a professor of corporate law at Boston College Law School, tells Fortune. “There’s very little legal basis for those kinds of claims. Once the money is handed over, that’s it.”
Throwing legal spaghetti at the wall
While nonprofits need to be mindful of how they treat donors if they want the checks to keep flowing, they have no shareholders with an economic interest that can be damaged.
Legally speaking, the responsibility to litigate on behalf of the public falls to the authorities—typically the attorney general of a state.
“If you donate to a charity, you don’t have a lot of recourse to later sue. U.S. law is not very favorable to donors in that regard,” Luís Calderón Gómez, a tax law specialist and assistant professor at Yeshiva University’s Cardozo Law School, tells Fortune.
He believes Musk’s chances may improve now that his team has shifted its legal strategy, increasing the number of alleged offenses to 14 from just four—even if the crux of the matter hasn’t changed.
Accusing OpenAI of everything from fraud and racketeering to false advertising and unjust enrichment may look like the legal equivalent of throwing spaghetti at a wall, hoping one of the charges somehow sticks.
Yet Calderón Gómez believes his case isn’t entirely without merit, given that courts may not look kindly on OpenAI’s brazen shedding of its nonprofit shell.
Musk has alleged the existence of a “Founding Agreement”, arranged with CEO Sam Altman, that expressly prohibits this eventuality. He has, however, failed to produce it, arguing instead it was reflected in the nonprofit’s December 2015 Certificate of Incorporation sufficiently to prove his point.
OpenAI moves to dismiss Musk lawsuit
“If he had that [Founding Agreement], he’d have a very strong case,” says Calderón Gómez. He believes Musk would be better off framing his dealings less as a donor and more as a co-founder who signed a binding document.
The burden of proof, however, lies with Musk—and demonstrating that OpenAI had planned to defraud him at the time of its founding will be challenging in the absence of clear evidence.
Listing no less than 82 separate legal precedents to back up its argument, lawyers for Altman’s company argued on Wednesday that Musk didn’t have a leg on which to stand.
“What’s new is that the carcass of the ‘Founding Agreement’ (now lowercased and shunted to the back of the complaint) is larded with allegations of fraud, racketeering, and false advertising,” OpenAI’s lawyers wrote dismissively. “But Musk offers neither the factual nor the legal scaffolding needed to sustain his claims.”
OpenAI’s legal team moved this week to dismiss the lawsuit outright, claiming Musk’s second attempt to drag the ChatGPT creator to court was just dressed up in even more hysterical language to deflect from its lack of substance.
Musk’s legal team at Toberoff & Associates didn’t respond to a request from Fortune for comment.
OpenAI among most valuable privately held companies
OpenAI has been increasingly candid about its plans to become a normal for-profit corporation, telling staff this will likely happen sometime next year.
At present, the company has no profits to distribute, and in fact, it continues to lose money due to the exorbitant costs of training and refining its models—by last account, $5 billion in red ink is expected for this year.
Moreover, a string of high-profile exits have occurred over the past six months, leaving CEO Sam Altman as one of only three founding team members left.
That hasn’t stopped investors from falling over themselves to buy shares in its operating company, which is titularly controlled by the nonprofit. Despite profits being capped, the equity it just raised valued it at $157 billion, making OpenAI one of the most valuable privately owned companies in the world.
Musk turns against his own creation
Musk, who left the board in 2018 and stopped donating entirely two years later, has meanwhile been forced to watch the success from the sidelines—a success he no longer could claim as his own.
Initially he still seemed like the proud parent. Days after ChatGPT launched at the end of 2022, he used his social media platform to draw attention to the invention and even chastised the New York Times twice for failing to cover it.
In the subsequent months, however, his tone changed dramatically as OpenAI grabbed headlines and triggered an explosion of interest in AI.
After it was clear ChatGPT would soon become the fastest-growing app in history, Musk began to speak out publicly against Altman’s research outfit, saying it had become the exact opposite of what he intended.
By May of 2023, it was clear he had an ax to grind.
‘I am the reason OpenAI exists’
That month, Musk told CNBC he had effectively wound up donating to a charity “to save the Amazon rainforest, and instead they became a lumber company, and chopped down the forest and sold it for money.”
At the time Musk was frustrated investors weren’t rewarding Tesla’s stock price for its own AI endeavors. He felt viewers ought to know Tesla was also on the cusp of its own ChatGPT moment once his Teslas could drive themselves without human supervision, a feat he qualified as “baby AGI”.
Musk wasn’t about to let OpenAI take all the credit two years after it cashed his last donation. “I am the reason OpenAI exists,” he said in the interview. By then Musk had already revealed plans to launch his own ChatGPT competitor, xAI, an idea that would eventually become a reality that July.
Emails reveal Musk wanted to seize control
In February this year, Musk finally sued the company, claiming it had broken its word to remain a nonprofit.
Shortly thereafter, OpenAI produced evidence showing Musk was well aware of this likelihood in late 2017, supported it himself, and only broke with the organization after he was not allowed to run it as CEO or subsume it into Tesla.
“Elon wanted majority equity, initial board control, and to be CEO,” the company revealed, sharing emails exchanged at the time. “We couldn’t agree to terms on a for-profit with Elon, because we felt it was against the mission for any individual to have absolute control over OpenAI. He then suggested instead merging OpenAI into Tesla.”
Exactly one day before OpenAI’s scheduled hearing over its first motion to dismiss, the law firm Irell & Manella, which represented Musk at the time, informed California’s Superior Court that it was withdrawing the lawsuit—with no explanation given.
Musk filed a second lawsuit in August, but nothing that OpenAI has seen has since changed its mind. “Elon’s recycled complaint is without merit and his prior emails continue to speak for themselves,” OpenAI told Fortune in a statement.
Courts likely won’t indulge a legal fishing expedition
Boston College Law School’s Quinn argues the Amazon rainforest analogy doesn’t constitute fraud, so long as Altman wasn’t actively shopping for chainsaws at the time. Transforming into a logging company several years after Musk was already out the door may be a morally questionable decision, but it isn’t an illegal one.
Barring a minor miracle, Quinn expects Musk’s case to be thrown out at the first opportunity. Musk might refile again since he has an effectively infinite ability to keep funding lawsuits, but eventually, a judge would sanction him.
“If there is no factual basis for the allegations, the court is not going to open its doors to say ‘anyone who wants to sue anyone else, just come on down here, put some words on paper, we’ll place unnecessary costs on defendants just for you to engage in a fishing expedition to come up with some damaging stuff,” Quinn says. “Courts are for good reason generally unwilling to allow plaintiffs to bootstrap lawsuits.”
Potential California lawsuit seen as having better odds
Musk can, however, take heart in knowing that at least one consumer advocacy group shares his frustration, even if for reasons other than personal.
The nonpartisan, Washington, D.C.-based Public Citizen has filed a complaint against Altman’s company with California attorney general Rob Bonta over its for-profit transformation.
Co-president Robert Weissman wrote the state AG “should insist that an OpenAI conversion to for-profit status reserve for humanity the right to any OpenAI invention of ‘artificial general intelligence’.”
In addition, Public Citizen wants OpenAI to pay roughly equivalent the value extracted from the nonprofit and handed to shareholders—to endow a new independent foundation for AI safety.
That alone should cost the new for-profit tens of billions of dollars, it estimates, something that could serve as a consolation to Musk.
When reached by Fortune, Bonta’s office wouldn’t say whether this has resulted in any enforcement action. “To protect the integrity of our investigations, we’re unable to comment on, even to confirm or deny, a potential or ongoing investigation,” it said.
Yeshiva University’s Calderón Gómez believes a state-launched case would have a far better outlook for success than Musk’s private lawsuit.
“If I were in the California AG’s office, I would probably sue,” he says. “There’s enough facts here that make me believe this hasn’t been operated as a nonprofit for a while now.”
