Ex-OpenAI Researcher Accuses CEO of Misleading Staff in Court Filing
A former OpenAI researcher has stepped into the legal fray against CEO Sam Altman, accusing him in a court filing of misleading employees about restrictive contracts and using the company's founding charter as a public-relations shield. Todor Markov, who left OpenAI for rival AI firm Anthropic, submitted a declaration as part of an amicus brief signed by 11 other former staffers, according to documents reviewed by this outlet.
Markov's central grievance centers on a non-disparagement clause that OpenAI required departing employees to sign to retain their vested equity in the multi-billion-dollar company. The clause, which barred ex-staff from criticizing OpenAI, came to light last year when Markov discovered its existence. During an all-hands meeting, Altman denied any knowledge of the provision, but that denial unraveled hours later when Vox published leaked documents showing the CEO had personally signed off on it.
In his declaration, Markov wrote that the episode convinced him Altman “was a person of low integrity who had directly lied to employees” about the agreements. He added that the incident led him to believe the CEO was “very likely lying to employees about a number of other important topics,” including OpenAI's commitment to developing safe artificial general intelligence (AGI).
Charter Called a 'Smokescreen'
Markov's filing takes particular aim at OpenAI's charter, which pledges to use any influence over AGI's deployment for the benefit of all and to avoid harm or undue concentration of power. The researcher said the charter was “positioned as the foundational document guiding all of our strategic decisions” but ultimately proved hollow. “I realized the Charter had been used as a smokescreen,” he wrote, “something to attract and retain idealistic talent while providing no real check on OpenAI's growth and its pursuit of AGI.”
The legal filing aligns Markov with Elon Musk, who has been vocal in his criticism of Altman and OpenAI's shift toward a for-profit structure. Musk, a cofounder of OpenAI who left the company years ago, has argued that the charter and mission were used as a facade to manipulate workers and the public. Markov, however, is not seeking any financial gain or acquisition, unlike Musk, and appears focused solely on airing his concerns.
The amicus brief, which includes statements from a dozen former employees, adds to the mounting pressure on Altman as he navigates a complex restructuring that would convert OpenAI's nonprofit arm into a for-profit entity. Critics argue that such a move contradicts the original mission, while supporters say it is necessary to fund the massive compute resources required for AGI development.
OpenAI has not yet responded to the latest filing, but the company has previously defended its practices, stating that non-disparagement clauses are common in the industry and that employee equity is contingent on standard legal agreements. The case is being watched closely by AI policy observers, as it could set a precedent for how AI companies balance trade secrets, employee rights, and public accountability.
For now, Markov's declaration stands as a rare insider account of the tensions between OpenAI's lofty ideals and its operational realities, offering a glimpse into the internal disillusionment that has led several key researchers to exit the company in recent years.
