THE FIRST ROUND of the legal battle between Elon Musk and Sam Altman over artificial intelligence technology is now underway.

To recap: On Feb. 29, Musk sued Altman and OpenAI Inc., the developers of ChatGPT, in San Francisco Superior Court, claiming that OpenAI had veered from the humanitarian goals enshrined in its nonprofit corporate charter. Musk’s complaint portrayed the fight in grand terms — some would say grandiose terms — as a fight over the future of humanity.

According to the complaint, Musk, Altman and Greg Brockman co-founded OpenAI to as a foil to, and counterweight for, the generative artificial intelligence products being developed at Google. The co-founders allegedly entered a so-called “Founding Agreement” that set forth their agreement that OpenAI would pursue artificial general intelligence as a nonprofit corporation and make its technology available on an open source basis for the benefit of humanity.

Artificial general intelligence, or AGI, describes technology that creates artificial intelligence that is able to equal or exceed humans in virtually all areas of human endeavor.

Musk claimed that under the leadership of Altman and Brockman, OpenAI has gone off course and, in return for billions in funding, let Microsoft Corporation get its hands on the crown jewels — the technology that created ChatGPT-4. According to Musk, this was a breach of the Founding Agreement.

He asked the San Francisco court to rectify the situation by ordering Altman, Brockman and OpenAI to comply with the agreement. He also asked the court to make a series of determinations about the state and status of OpenAI’s pursuit of AGI.

OpenAI hadn’t stepped into the ring with Musk for a minute before it took a shot at a knockout.

In a demurrer and a motion to strike filed March 27, it said that Musk’s argument had a fatal flaw: the Founding Agreement was “make-believe.” In other words, the agreement did not exist and Musk hadn’t shown that there was any contract at all — written, oral, or implied — that supported his claim. In a pithy putdown, the company said, “Musk cannot bring suit on an agreement that he cannot even plead into existence.”

The pleadings also savaged Musk’s other claims and set out a counter-narrative that described Musk as a disgruntled competitor in the AI space who was angry, not over the fate of humanity, but because Altman and Brockman had blocked his plan to take control of OpenAI himself.

Musk responds

Musk responded to the demurrer and the motion to strike on Wednesday.

A demurrer is a legal challenge to the legal sufficiency of a plaintiff’s complaint. In essence, it asks the court to rule that even if everything that the plaintiff says is found to be true after a trial, it still wouldn’t be enough for plaintiff to win. It is a way for the court to weed out, at the beginning of the litigation, suits where the facts — as stated by plaintiff — don’t amount to a valid legal claim.

A demurrer is generally a relatively low bar for a plaintiff to clear because the court is required to accept the plaintiff’s version of the facts as long as they are not just unsupported conclusions.

Musk’s response to the demurrer went directly at OpenAI’s lead argument — that Musk had not set out facts that supported his contention that there was a Founding Agreement.

Musk pointed out that his complaint said the Founding Agreement was “memorialized in OpenAI’s written Articles of Incorporation and in written communications between Plaintiff and Defendants.”

The OpenAI logo appears in a photo from its website homepage. The company behind the artificial intelligence software ChatGPT has filed a motion to strike portions of a lawsuit brought against it by former cofounder Elon Musk. (Glenn Gehlke/Bay City News)

In this respect, Musk was saying that a contract can exist in different documents that are all interpreted together to give effect to the parties’ intentions.

Musk put heavy weight on the documents filed in Delaware when OpenAI Inc was founded as a nonprofit corporation. Musk quoted sections of those documents that said:

• OpenAI “shall be a nonprofit corporation organized exclusively for charitable and/or educational purposes.”

• The company’s purpose “is to provide funding for research, development and distribution of technology related to artificial intelligence” that “will benefit the public.”

• The company “is not organized for the private gain of any person”

• Company property “is irrevocably dedicated to” such nonprofit purposes

• “The technology would be owned by the foundation and used ‘for the good of the world.’”

• OpenAI “will seek to open source technology for the public benefit when applicable.”

Musk cited a prior case where the terms of a corporate charter served as the basis for a breach of contract claim. He also said that whether or not there was a contract was disputed fact, something to be decided only later in the case after the parties had taken discovery of each other and presented their competing allegations to the judge or jury.

A belt and suspenders maneuver

In addition to the demurrer, OpenAI filed a motion to strike portions of Musk’s complaint that asked the court to make determinations about the state of development of OpenAI’s technology and then use those determinations “to enter an exorbitant and impractical order of specific performance.” Specific performance is a legal remedy that orders a party to carry out a contract as written.

The motion to strike was something of a belt and suspenders maneuver for OpenAI because if Musk’s complaint is found insufficient in its entirety as the demurrer argues, it would be unnecessary to strike any part of it. However, if some or all of the complaint survives the demurrer, the defendants want the surviving part to be stripped of the portions that seek the most far-reaching relief.

In the motion, the defendants zero in on Musk’s contention that OpenAI is in the process of developing a new AI product called Q* that is more powerful than ChatGPT. Musk asks the court to determine whether Q* constitutes AGI. Musk claims that question to be of profound importance because OpenAI has only licensed Microsoft its pre-AGI technologies; if Q* is AGI, it belongs to humanity, according to Musk, not a for-profit corporation.

The defendants argue that what Musk is asking the court to do is not a proper role for a court. They argue that even though “Musk does not even know what Q* is,” he wants the court to wade into the swamp of determining whether an in-development technology amounts to AGI. They say that would amount to a hypothetical and formless inquiry, one that is inappropriate for a court that is charged with deciding actual controversies, not speculations.

Musk’s response says that calling the technology “in-development” puts the rabbit in the hat. He says he believes that OpenAI’s board — a board he contends that Altman has hand-selected to be “captured, conflicted, and compliant” — will never determine that the technology is AGI because the defendants want to be able to continue to provide it to Microsoft.

If a court cannot review it because it is “in development,” it will always be in development.

The demurrer and motion to strike raise many other arguments, some outside the usual scope of such filings.

Not yet down for the count

Where this will go from here cannot yet be determined, but a complete knockout blow to Musk seems unlikely.

In many disputes of this complexity, a court will use the filings as an opportunity to narrow and simplify the case by paring down some of the claims yet letting the plaintiff proceed as to the remainder. And even if a complaint is completely dismissed, the court may give the plaintiff an opportunity to amend and refile.

One wildcard in this situation relates to OpenAI’s status as a nonprofit corporation.

Separate from Musk’s contract and his other private claims, there could be a nonprofit corporate law question of whether OpenAI is operating within the limitations of its charter and applicable nonprofit law. In some instances, corporate law gives the state attorney general responsibility for making sure that corporations with a charitable mission do not ignore their corporate purposes.

OpenAI was formed in Delaware so questions like that might be for that state, although the parties have filed papers that show that on Aug. 28, 2017, OpenAI, Inc. was registered with the Registry of Charitable Trusts maintained by California’s Attorney General. The registration requirement, subject to some exemptions, applies to charitable corporations “holding assets for charitable purposes or doing business in the State of California.”

When the act applies, California Government Code 12588 says, in part, that “The Attorney General may investigate transactions and relationships of corporations and trustees subject to this article for the purpose of ascertaining whether or not the purposes of the corporation or trust are being carried out in accordance with the terms and provisions of the articles of incorporation or other instrument.”

According to court records, California’s Attorney General Rob Bonta has not sought, at least at this point, to be involved in Musk’s pending lawsuit.

Joe Dworetzky is a second career journalist. He practiced law in Philadelphia for more than 35 years, representing private and governmental clients in commercial litigation and insolvency proceedings. Joe served as City Solicitor for the City of Philadelphia under Mayor Ed Rendell and from 2009 to 2013 was one of five members of the Philadelphia School Reform Commission with responsibility for managing the city’s 250 public schools. He moved to San Francisco in 2011 and began writing fiction and pursuing a lifelong interest in editorial cartooning. Joe earned a Master’s in Journalism from Stanford University in 2020. He covers Legal Affairs and writes long form Investigative stories. His occasional cartooning can be seen in Bay Area Sketchbook. Joe encourages readers to email him story ideas and leads at joe.dworetzky@baycitynews.com.