Sam Altman

OpenAI trained its commercial models on copyrighted material without consent or compensation — building an $852 billion enterprise on content it did not pay for. Books1 and Books2, the datasets used for GPT-3 training, contained over 100,000 copyrighted books. When this became public in May 2024, OpenAI destroyed both datasets — after litigation had already been filed. Courts treat post-litigation evidence destruction as spoliation. For GPT-4, OpenAI transcribed more than one million hours of YouTube videos for training data — acknowledged in court filings. More than a dozen major lawsuits are pending, including suits from the New York Times, the Authors Guild (representing George R.R. Martin, John Grisham, Jodi Picoult, and others), eight major newspapers, and Canadian national outlets. No settlements have been disclosed. The copyright violation is embedded in every response ChatGPT has ever produced.

• —Books1 and Books2 datasets: 100,000+ copyrighted books used for GPT-3 training. Disclosed publicly May 2024. Datasets destroyed after litigation commenced.
• —GPT-4 training: 1M+ hours of YouTube video transcribed — acknowledged in OpenAI court filings.
• —Active lawsuits (partial): Sarah Silverman et al. (July 2023); Authors Guild / Martin, Grisham, Picoult, Franzen (September 2023); The New York Times (December 2023); The Intercept, Raw Story (February 2024); 8 newspapers including Chicago Tribune, Denver Post (April 2024); Toronto Star, Globe and Mail, CBC, Postmedia (November 2024); Ziff Davis (April 2025).
• —Post-litigation destruction of Books1 and Books2 is the conduct courts classify as spoliation — raising an inference that the destroyed evidence would have been damaging.
• —OpenAI publicly lobbied for "fair use" exemptions to copyright law while simultaneously destroying the datasets that would have documented the scope of the violation.

On November 17, 2023, OpenAI's safety-focused board — Helen Toner (Georgetown AI researcher), Ilya Sutskever (co-founder), Adam D'Angelo, Tasha McCauley — fired Sam Altman as CEO, stating he "was not consistently candid in his communications with the board." Within hours, Microsoft (27% stakeholder) objected. Investors protested. 738 of OpenAI's 770 employees signed an open letter threatening to resign and join Microsoft if Altman wasn't reinstated and the board didn't resign. Five days later, Altman was reinstated. All safety-focused board members departed. The new board: Bret Taylor (chairman, former Salesforce co-CEO) and Lawrence Summers (former Treasury Secretary) — commercial and financial figures, not safety researchers. The governance structure created to check Altman was replaced by governance that would not. The aftermath: Ilya Sutskever resigned in May 2024, citing safety concerns. The superalignment project — designed to develop techniques for governing superintelligent AI before it arrived — was dissolved 18 months after launch. Approximately half of OpenAI's AI safety researchers left the company throughout 2024, publicly citing the organization's "deprioritization of safety goals."

• —Board firing (November 17, 2023): stated reason — Altman "was not consistently candid in his communications with the board."
• —738 of 770 OpenAI employees signed open letter threatening mass resignation if Altman wasn't reinstated.
• —Reinstatement (November 21, 2023): 5 days after firing. All safety-focused board members replaced.
• —New board: Bret Taylor (former Salesforce co-CEO), Lawrence Summers (former Treasury Secretary). Safety expertise: zero.
• —May 2024: Ilya Sutskever resigned. Superalignment project — created to solve AI oversight before superintelligence — simultaneously dissolved.
• —Jan Leike (superalignment co-lead) departed simultaneously, posting publicly about safety deprioritization.
• —~50% of OpenAI's AI safety researchers left throughout 2024, citing "deprioritization of safety goals."
• —OpenAI now races toward AGI without the safety oversight structure it publicly committed to building.

OpenAI was incorporated in 2015 as a nonprofit corporation with assets legally dedicated to its public benefit mission. In October 2025, Altman completed the conversion of OpenAI into a for-profit Public Benefit Corporation. The nonprofit foundation — which used to own the entire enterprise — now holds a 26% stake. At the $852 billion valuation (April 2026), that 26% is worth approximately $221 billion. The foundation used to own 100% of an enterprise now worth $852 billion. The founders, investors, and employees capture the difference. Multiple critics — including legal scholars and Elon Musk's litigation team — characterized the transaction as an illegal diversion of charitable assets. The California and Delaware Attorneys General reviewed the transaction. The FTC opened an investigation in July 2023 examining "circular" spending arrangements between OpenAI and Microsoft. No criminal charges have been filed. The conversion is complete. Sam Altman's personal equity stake — which was zero under the nonprofit structure — has not been publicly disclosed. The company is valued at $852 billion.

• —2015: OpenAI incorporated as nonprofit. Assets legally dedicated to mission of benefiting humanity. No private inurement allowed.
• —December 2024: Restructuring announced — conversion to Public Benefit Corporation.
• —October 28, 2025: Conversion completed. OpenAI Group PBC formally created.
• —Post-conversion ownership: OpenAI Foundation (nonprofit): 26%. Microsoft: 27%. Investors + employees: 47%.
• —Valuations at time of conversion: ~$500B (employee share sale); grew to $852B by April 2026.
• —Legal letter "Not For Private Gain": argued the nonprofit was legally obligated to retain control of assets dedicated to its public mission.
• —Elon Musk litigation (revised April 2026): sought reversal of the corporate restructuring on grounds of illegal diversion of charitable assets.
• —Sam Altman's personal equity: was zero under nonprofit structure. Post-conversion equity stake has not been publicly disclosed. The company is worth $852B.

Sam Altman co-founded and chairs Tools for Humanity, which operates the "World" (formerly Worldcoin) project. The model: send physical "Orb" devices to low-income countries, offer cryptocurrency tokens in exchange for iris scans, and build a global biometric identity database. The iris is the most distinctive and un-changeable piece of biometric data a human possesses. The target populations — in Africa, Southeast Asia, and Latin America — have limited income and limited legal recourse when data is misused. The iris data cannot be revoked if it is breached or sold. By September 2025, the project had 15 million iris-verified users. The regulatory verdict from a dozen countries was consistent: illegal. Kenya (August 2023): suspended enrollment; May 2025 ordered full deletion of all biometric data collected. Spain (March 2024): ordered halt to biometric data collection. Portugal (March 2024): temporary suspension. Hong Kong (January 2024): found violations of the Personal Data Ordinance. Philippines (October 2025): cease and desist for multiple data privacy violations. Thailand (November 2025): ordered halt and deletion of data from 1.2 million users. France, UK, Bavaria, South Korea, Brazil, India, Indonesia all opened investigations or suspended operations.

• —Business model: iris scan → cryptocurrency token. Permanent, irrevocable biometric data exchanged for speculative digital currency.
• —Scale: 33 million World App users (September 2025); 15 million iris-verified.
• —Geographic concentration: Africa, Southeast Asia, Latin America — low-income populations with limited legal recourse.
• —Kenya (August 2023): suspended enrollment over "security, privacy and financial concerns." May 2025: ordered full deletion of all biometric data.
• —Spain (March 2024): ordered halt to biometric data collection.
• —Philippines (October 2025): cease and desist for multiple data privacy violations.
• —Thailand (November 2025): ordered halt; deletion ordered for 1.2 million users' data.
• —Additional enforcement/investigations: Hong Kong, Portugal, France, UK, Bavaria (Germany), South Korea, Brazil, India, Indonesia.
• —Altman continues operating Worldcoin/World in jurisdictions where regulators have not yet acted.

Sam Altman has pursued a consistent strategy to prevent regulation of AI during the period of maximum commercial expansion. In December 2024, at the precise moment OpenAI was completing its nonprofit-to-PBC conversion — requiring federal regulatory non-interference — Altman donated $1 million to Donald Trump's inaugural fund. OpenAI actively opposed California's AI safety legislation (SB 1047), arguing it "encroaches on a more competent federal government" — a federal government that has no AI safety regulation in place. The effect of this position: oppose regulation where it exists (states); advocate for regulation where it doesn't (federal). The gap between those two positions is the window during which OpenAI deploys its products at scale. In July 2023, the FTC opened a civil investigation into OpenAI's data security and privacy practices, flagging "circular" spending arrangements between OpenAI and Microsoft as potentially anticompetitive. Altman simultaneously testified before Congress in May 2023 calling for AI regulation — while opposing every state-level bill that would have applied immediately.

• —$1M to Trump Inaugural Fund (December 2024): timed to OpenAI's nonprofit-to-PBC conversion completion; also concurrent with ongoing federal contract negotiations.
• —California SB 1047 opposition: OpenAI argued the bill "encroaches on a more competent federal government" — a federal government with zero AI safety statutes in force.
• —Pattern: oppose state regulation (exists now, would apply now); advocate for federal regulation (does not exist; creates regulatory vacuum).
• —FTC civil investigative demand (July 2023): examined data security/privacy practices; flagged "circular" OpenAI-Microsoft spending arrangements.
• —Congressional testimony (May 2023): publicly called for AI regulation while simultaneously opposing every actionable bill at the state level.
• —The regulatory vacuum during 2023–2025 is the period in which ChatGPT went from zero to 400M+ weekly users and OpenAI's valuation went from ~$30B to $852B.