- OpenAI once stated (in 2023) that its goal was to build AGI “safely for the benefit of humanity, unconstrained by a need to generate financial return.”
- In its IRS 990 filing for 2024 (filed in late 2025), the mission was changed to “ensuring AGI benefits all of humanity” — dropping the word “safely” and the commitment to be “unconstrained by profit.”
- The change occurred alongside the transition from a non-profit model to a for-profit structure.
- In October 2025, OpenAI reached an agreement with the Attorneys General of California and Delaware to restructure into two entities: OpenAI Foundation (non-profit) and OpenAI Group (public benefit corporation).
- The Foundation owns about 26% of the shares; Microsoft holds 27% after investing a total of $13.8 billion; the rest belongs to employees and other investors.
- In late 2024, OpenAI raised $6.6 billion; after restructuring, SoftBank completed a $41 billion investment and is negotiating an additional $30 billion.
- The company’s valuation exceeded $500 billion (up from $300 billion in March 2025).
- The article notes that OpenAI is facing multiple lawsuits related to product safety (including allegations of psychological manipulation, wrongful death, and negligence).
- Although there is a safety committee in the Foundation and a requirement for annual benefit reports, neither entity explicitly states “safety” in its mission; many members sit on both boards, raising concerns about independent oversight.
- The author suggests this is a “test” of AI governance: whether social priorities will be overshadowed by shareholder pressure and a potential IPO.
📌 In its IRS 990 filing for 2024 (filed in late 2025), OpenAI’s mission changed to “ensuring AGI benefits all of humanity” — dropping the word “safely” and the commitment to be “unconstrained by profit.” OpenAI’s removal of the word “safety” from its mission runs parallel to its restructuring to attract massive capital and a valuation exceeding $500 billion. Despite still claiming to value safety, the change in wording and control (the Foundation holds only 26%) raises questions about accountability. This is a critical test: Will AI be governed for social benefit or for shareholder profit?
