Nonprofits that challenged OpenAI were hit with subpoenas
▼ Summary
– OpenAI has subpoenaed multiple nonprofits including The Midas Project, demanding extensive donor information and communications about OpenAI’s governance and restructuring.
– Recipients and legal experts describe the subpoenas as overly broad “fishing expeditions” that appear designed to intimidate critics rather than gather relevant lawsuit evidence.
– The subpoenas impose significant financial burdens through legal fees and have made some organizations uninsurable, effectively constraining critical speech.
– OpenAI claims the subpoenas help defend against Elon Musk’s lawsuit, but critics note they target policy work unrelated to the case, including state AI safety legislation.
– Internal OpenAI employees have publicly criticized the tactics, with one mission alignment lead calling them “oppressive” and warning they turn OpenAI into “a frightening power.”
A late evening text message alerted Tyler Johnston that a process server was at his door, marking the start of a legal confrontation between his tiny nonprofit, The Midas Project, and the AI giant OpenAI. Johnston’s organization, which scrutinizes AI firms for ethical compliance and transparency, had published a detailed report and circulated an open letter questioning OpenAI’s shift to a for-profit model. Now, OpenAI responded not with dialogue but with subpoenas, delivered both to his group and to him personally, suggesting his work was secretly directed by Elon Musk.
While traveling in California, Johnston learned that a firm called Smoking Gun Investigations, LLC, acting for OpenAI, had attempted physical service before sending two 15-page subpoenas by email. Though he anticipated legal pressure, the sweeping demands took him aback. OpenAI didn’t just ask whether Musk funded The Midas Project; it sought every donor’s identity and contribution details, plus all documents and communications concerning OpenAI’s governance and structural changes. Johnston noted that public tax forms showed his nonprofit’s modest sub-$75,000 annual revenue, quipping that any Musk backing would have been exceptionally frugal.
Johnston’s situation is far from isolated. In recent months, at least seven nonprofits critical of OpenAI’s restructuring have been hit with similar subpoenas, including the San Francisco Foundation, Encode, Ekō, the Future of Life Institute, and others. Recipients and legal experts argue these legal demands function less as legitimate discovery and more as intimidation, imposing heavy administrative and financial burdens. The controversy has spilled into public view, with even an OpenAI mission alignment lead, Joshua Achiam, posting on X that the tactics “don’t seem great” and risk making the company “a frightening power.”
For Johnston, compliance would have meant surrendering confidential material, including sensitive discussions with former OpenAI staff and private data on open letter signatories. James Grimmelmann, a Cornell Law School professor, finds it “really hard” to see the legal relevance of probing Musk’s potential influence, calling the subpoenas “chilling” and “oppressive.” He points out that the extensive document searches and legal responses could drain a small nonprofit’s resources, comparing the strategy to Musk’s own litigious campaigns against critics.
Though Johnston secured legal counsel and learned he need not fully comply, the ordeal has had real consequences. Insurers denied his organization coverage, citing the OpenAI-Musk dispute, a constraint he calls another form of limiting free expression. Meanwhile, a judge who initially permitted OpenAI’s discovery efforts indicated the court was reconsidering after seeing the “scope” of the demands.
Nathan Calvin, general counsel at the policy nonprofit Encode, was served during dinner by a sheriff’s deputy. He found one request particularly troubling: OpenAI demanded all documents about California’s AI safety law, SB 53, which Encode helped draft and OpenAI publicly opposed. Calvin questioned what the company hoped to gain by accessing Encode’s lobbying communications and contacts. After Encode went public with the subpoena, OpenAI CSO Jason Kwon posted that there was “quite a lot more to the story,” emphasizing their legal defense against Musk and noting Encode had filed an amicus brief against OpenAI’s restructuring.
Other subpoenas reportedly sought documents on additional California bills, such as SB 1047 and AB 501, indicating a pattern of targeting policy-focused critics. Sacha Haworth of the Tech Oversight Project, which co-released The OpenAI Files, accused OpenAI of deploying “lawfare” and “paranoid accusations” to intimidate advocates, following a Big Tech playbook of leveraging legal and political power.
Max Tegmark of the Future of Life Institute confirmed receiving a similarly broad subpoena, while Judith Bell of the San Francisco Foundation described OpenAI’s demands as a “fishing expedition” into their strategic work. Bell emphasized that the real issue isn’t the subpoenas but OpenAI’s restructuring and the fate of charitable assets she says belong to the public. By focusing on legal tactics, she noted, attention shifts away from the core questions of equity, governance, and oversight in OpenAI’s new for-profit era.
(Source: The Verge)
