SRC - OpenAI
Folder: 05 - SOURCE MATERIAL Supporting: OpenAI
Corporate Structure and History
OpenAI — Wikipedia https://en.wikipedia.org/wiki/OpenAI Confirms: Founded 2015 as nonprofit. For-profit subsidiary 2019. Restructured to Delaware PBC October 2025. Microsoft investment 50 billion February 2026. Safety team dissolution May 2024. Ilya Sutskever and Jan Leike departures confirmed.
Sam Altman Firing — Wikipedia https://en.wikipedia.org/wiki/Removal_of_Sam_Altman_from_OpenAI Confirms: Fired November 2023 for “lack of candour.” 743 of 770 employees threatened to quit if not reinstated. Reinstated five days later. March 2024 investigation concluded behaviour “did not mandate removal.” Helen Toner explanation confirmed — Altman withheld information about ChatGPT release and safety processes.
Helen Toner — CNBC Interview May 29, 2024 https://www.cnbc.com/2024/05/29/former-openai-board-member-explains-why-ceo-sam-altman-was-fired.html Confirms: board learned about ChatGPT launch from Twitter. Altman did not disclose ownership of OpenAI startup fund. Safety process misrepresentations confirmed.
The Microsoft Deal — AGI Clause
Secret AGI clause in Microsoft agreement: AGI defined as point where system generates “maximum total profits to which its earliest investors are entitled.” Referenced in multiple sources including Time Magazine analysis and restructuring documentation.
Time Magazine — May 30, 2024 “How Anthropic Designed Itself to Avoid OpenAI’s Mistakes” https://time.com/6983420/anthropic-structure-openai-incentives/ Confirms: OpenAI structural evolution and incentive misalignment documented as context for Anthropic’s founding.
October 2025 Restructuring: AGI now verified by independent expert panel. Microsoft access secured through 2032. Wikipedia — OpenAI above.
Safety Team Departure
Wikipedia — OpenAI “Throughout 2024, roughly half of then-employed AI safety researchers left the company.” Ilya Sutskever and Jan Leike co-leaders both left. Superalignment team disbanded May 2024 after promising 20% compute resources — team denied receiving anything close to 20%.
Jan Leike resignation statement: Loss of belief that “when OpenAI says it’s going to do something or says that it values something, that that is actually true.” Widely reported May 2024. Multiple sources confirmed.
Privacy and Data Collection
OpenAI Privacy Policy Updated February 6, 2026 https://openai.com/policies/row-privacy-policy/ https://openai.com/policies/privacy-policy/ Confirms: Account information, user content, communications, voice data, images, and documents collected. Consumer ChatGPT conversations used for training by default. Opt-out available. Enterprise and API excluded from training by default.
OpenAI Consumer Privacy Page https://openai.com/consumer-privacy/ Official controls documentation.
Lexology — OpenAI Privacy Update April 14, 2025 https://www.lexology.com/library/detail.aspx?g=a9b322ad-aebc-4d91-83cf-4b70f4911b34 Confirms: Three primary training data sources — account information, user content, communications. Late 2024 policy update documented.
The Court Order — 400 Million Users
Nelson Mullins — July 10, 2025 “From Copyright Case to AI Data Crisis” https://www.nelsonmullins.com/insights/blogs/corporate-governance-insights/all/from-copyright-case-to-ai-data-crisis-how-the-new-york-times-v-openai-reshapes-companies-data-governance-and-ediscovery-strategy Key findings:
- May 13, 2025: Judge Wang issued preservation order requiring OpenAI to retain ALL ChatGPT conversation logs — affecting 400+ million users
- First court-mandated mass preservation of AI content at this scale
- ChatGPT Free, Plus, Pro, and Team subscriptions all affected
- API users without Zero Data Retention agreements affected
- Conversations stored in “secured system” accessed only by “small, audited OpenAI legal security team”
- Sam Altman proposed “AI Privilege” to protect sensitive conversations — no such privilege recognised in US law
OpenAI Official Response — June 5, 2025 https://openai.com/new-york-times/ Confirms: preservation order scope. New York Times demanded 1.4 billion private ChatGPT conversations be turned over — May 2025. OpenAI pushing back continuously. Court required indefinite retention of chats users chose to delete.
The August 2025 Privacy Incident
Wikipedia — OpenAI August 2025: Thousands of private ChatGPT conversations inadvertently exposed to public search engines. Experimental “share with search engines” feature malfunctioned. Personal details, names, locations, intimate topics appeared in search results. Feature permanently removed August 1, 2025. Sam Altman acknowledged the issue. Not a security breach — design flaw.
Copyright Lawsuits
New York Times v. OpenAI Filed December 2023. NYT demanded 1.4 billion private ChatGPT conversations — May 2025. OpenAI alleged NYT secretly deleted evidence of its own extensive use of OpenAI models internally. Case ongoing as of March 2026.
Fair Use Rulings — 2025 Two federal judges independently confirmed AI training is “highly transformative” and protected by fair use doctrine. Bartz v Anthropic — June 23, 2025. Kadrey v Meta Platforms — June 25, 2025. Referenced in OpenAI’s own legal documentation above.
The Verge / Multiple sources: Authors, news organisations, and media companies have filed multiple copyright infringement suits. All cases ongoing.
Elon Musk Lawsuit
Wikipedia — OpenAI April 9, 2024: OpenAI countersued Musk alleging “bad-faith tactics.” February 10, 2025: Musk consortium submitted 134 billion headed to jury trial April 2026.
Funding
Wikipedia — OpenAI 500 billion at restructuring. Amazon $50 billion February 2026. Microsoft threatened to sue over Azure exclusivity violation.
Saudi Arabia fundraising: Altman fundraising with Saudi Arabia for chip manufacturing capacity — alarmed safety researchers. Multiple sources including Bloomberg confirmed.
Still To Source
- Secret AGI clause in Microsoft agreement — primary document location needed investigate
- Full scope of government contracts beyond consumer use — OpenAI DoD relationships investigate
- Musk $134 billion lawsuit trial outcome — April 2026, monitor investigate
- Amazon deal Microsoft lawsuit status — monitor investigate
- Restructuring nonprofit equity arrangement — full terms not publicly disclosed #investigate
- ChatGPT preservation order — appeal outcome pending #investigate