OpenAI began as a nonprofit research lab with a public-facing promise: powerful AI should not be shaped only behind closed doors. That promise is now under renewed scrutiny because documents the organization previously said were available for public review are not being shared.
According to the source article, OpenAI’s reports to US tax authorities had long said that members of the public could review governing documents, financial statements, and conflict of interest rules. When WIRED requested those materials last month, OpenAI said the policy had changed.
What OpenAI Said It Would Share
OpenAI was launched in 2015 by wealthy tech entrepreneurs including Elon Musk. Its original nonprofit identity was tied to the idea that society and the public should have a role in the development of powerful AI, in contrast with large technology companies working privately.
That posture appeared in OpenAI’s annual IRS filings. For seven consecutive years, from its founding through 2022, OpenAI stated that its annual submissions and other files were available “upon request.”
The documents at issue are not minor public-relations materials. They include records that could help outsiders understand how the nonprofit is governed, how conflicts are handled, and how financial reporting is presented.
- Governing documents could show how board authority is structured.
- Conflict of interest rules could clarify how outside interests are managed.
- Financial statements could help explain what the nonprofit does and does not disclose.
The Policy Shift
When WIRED asked for the records, OpenAI did not provide the full set of materials described in the earlier filings. Instead, the company supplied a narrow financial statement that did not include most of its operations.
OpenAI spokesperson Niko Felix explained the company’s position this way: “We provide financial statements when requested,” adding, “OpenAI aligns our practices with industry standards, and since 2022 that includes not publicly distributing additional internal documents.”
That answer marks a clear break from the earlier public-facing language in the filings. The company’s IRS filing for 2023, which would reflect the changed policy, was not due until later this year, according to the source article.
Felix also said San Francisco-based OpenAI discloses all material required by the IRS and California’s attorney general. He said the company regularly publishes information about its research and safety work and makes research available through tools such as ChatGPT.
Why Governance Records Are Under Scrutiny
The request came after a high-profile crisis at OpenAI. In November, OpenAI’s board fired CEO Sam Altman, saying in a statement that he had endangered its mission to ensure AI “benefits all humanity.”
An employee and investor revolt soon pushed the board to reinstate Altman. Most of the board’s own members were then ejected, and a revised slate of directors vowed to review the crisis and make structural changes to regain stakeholder trust.
That sequence made OpenAI’s governing documents more important, not less. The records could help explain what authority the board has, what changes were made after Altman returned, and how the unusual structure of the organization operates in practice.
The source article notes that OpenAI’s original nonprofit organization and its board retain ultimate control of OpenAI’s activities and technology. That makes the nonprofit’s internal rules central to understanding how decisions are made, even as the company’s best-known products are associated with a wider set of operations.
Conflicts, Investments, and Board Power
OpenAI’s conflict of interest policy is another key missing document. Access to it could show how the board oversees Altman’s outside pursuits, which the source article says include personal investments in numerous startups pursuing AI projects and a nuclear reactor maker.
People involved in the situation but not authorized to discuss it said the overlap between Altman’s day job and personal projects played some role in board members’ distrust. In 2019, while Altman was leading the company, OpenAI signed a nonbinding letter of intent to buy $51 million of AI chips from Rain, a startup in which he has invested more than $1 million, WIRED reported last month.
OpenAI has not moved forward with that purchase. Felix said Altman is transparent with the board about his investments and follows a process for managing potential conflicts.
The public still cannot compare that description with the actual conflict rules OpenAI once said were available on request. That gap is why the withheld policy matters: it limits outsiders to company statements rather than the underlying standard the company says it applies.
A Broader Retreat From Openness
The dispute fits a wider pattern described in the source article. OpenAI once published extensive details about its AI inventions, but it has more recently protected the technical details and data behind ChatGPT.
The shift became especially visible after 2019, when the nonprofit created a for-profit subsidiary to house most AI development and attract outside investment. That move helped connect OpenAI’s future to Microsoft, one of the large technology companies it was originally founded to challenge.
OpenAI cofounder turned competitor Elon Musk criticized the company at a New York Times event in November, saying it should be called Super-Closed-Source-for-Maximum-Profit-AI.
The source article also explains that not all nonprofits make governing documents and internal policies public. Rick Cohen, chief operating and communications officer for National Council of Nonprofits, said, “It is not common for organizations to make their governing documents or internal policies public.”
But OpenAI’s case is different because it previously told the IRS that these materials were available “upon request.” The issue is not only what nonprofit rules require. It is whether a company built on public trust is still willing to honor the transparency it once advertised.