What 'The OpenAI Files' uncovers about Sam Altman—and company's governance
What's the story
A new investigative report, "The OpenAI Files," offers an unprecedented look into the inner workings of OpenAI, examining its leadership, structure, and evolution.
Produced by the Midas Project and Tech Oversight Project, and put together by Tyler Johnston, the analysis compiles public records, corporate disclosures, media reports, and legal filings.
It is touted as the most exhaustive collection of documented concerns related to governance practices and organizational culture at the Sam Altman-led AI firm.
Comprehensive overview
Report details OpenAI's transition from nonprofit to profit-driven entity
Spanning over 50 pages and more than 10,000 words, the interactive report details OpenAI's transformation from a nonprofit research lab to a profit-driven entity.
The report charts OpenAI's journey from a nonprofit research lab to a hybrid capped‑profit corporation, illustrating key turning points in governance and organizational culture.
It also highlights the safety concerns and potential conflicts of interest that have arisen during this transition.
Executive gains
Speculative look at Altman's investment portfolio included
One major focus of the report is the potential financial benefits that OpenAI executives and board members could reap from the company's success.
The report even includes a speculative look at CEO Altman's investment portfolio, which features companies like Retro Biosciences, Helion Energy, Reddit, Stripe, Rewind AI and Rain AI.
These firms are said to have current or future business ties with OpenAI through partnerships or acquisition talks.
Unbiased perspective
Report created without any funding from OpenAI competitors
Johnston, who is the executive director of The Midas Project, clarified that the report was created without any funding or support from OpenAI competitors like Elon Musk, xAI, Anthropic, Meta, Google and Microsoft.
He emphasized that this project is an archival effort to show how OpenAI has evolved over time.
"We're just putting that information in front of the reader and asking the reader to draw their own conclusions about what to make of it," Johnston told The Verge.