disinfozone/The_Info_Web
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
The_Info_Web/20 - ORGANIZATIONS/PRIVATE/OpenAI.md

29 KiB
Raw Permalink Blame History

category tags summary location created updated relations
Private Organization
Organization
OpenAI
ArtificialIntelligence
EffectiveAltruism
PeterThiel
ElonMusk
SiliconValley
OpenAI is an artificial-intelligence company founded as a nonprofit in 2015 with a billion-dollar pledge from backers including Elon Musk, Sam Altman, Reid Hoffman, and Peter Thiel, whose 2023 firing and reinstatement of CEO Sam Altman exposed the conflict between its effective-altruism-aligned safety board and its commercial expansion. San Francisco, California 2026-06-19 2026-10-15
type with reverse start role fn
employed_by Abdulrahman Alfozan true 2025-07 engineer; arrested October 9, 2026 12
type with reverse start role fn
owned MGX true 2025-10 held shares acquired in a secondary sale of about 6.6 billion dollars at a valuation of around 500 billion dollars 22
type with reverse start role fn
funded Andreessen Horowitz true 2026-03 named co-lead, per Bloomberg, of a roughly 10 billion dollar financing said to be set to close the following week 23
type with reverse start role fn
funded MGX true 2026-03 named co-lead, per Bloomberg, of a roughly 10 billion dollar financing said to be set to close the following week 23
type with start role fn
partner_of G42 2025-05 Stargate UAE, 1 GW cluster in Abu Dhabi, 200 MW expected to go live in 2026 19
type with reverse start role fn
director_of Paul Nakasone true 2024-06-13 board member; also named to the board's security and safety subcommittee 24
type with start role fn
partner_of Anduril Industries 2024-12-04 counter-drone models partnership 25
type with start role fn
partner_of Los Alamos 2025-01-30 o-series models on the Venado supercomputer, moved to a classified network 26
type with start end role fn
contractor_to Chief Digital and Artificial Intelligence Office 2025-06-16 2026-07 200 million dollar prototype other transaction agreement, estimated completion July 2026 28
type with start role fn
contractor_to General Services Administration 2025-08-06 OneGov agreement, ChatGPT Enterprise at 1 dollar per agency for one year 29

OpenAI is an artificial-intelligence research company founded in San Francisco on December 11, 2015 as a nonprofit laboratory whose stated mission was to ensure that artificial general intelligence benefits humanity. Its founding was announced with a one-billion-dollar funding pledge from Elon Musk, Sam Altman, Greg Brockman, Reid Hoffman, Jessica Livingston, Peter Thiel, Amazon Web Services, and Infosys, though tax filings later showed that only about 133 million dollars had actually been received by 2021. The organization grew out of the same AI-existential-risk concern that animates the Rationalist Community and Effective Altruism movements, and its 2023 governance crisis exposed the fault line between that safety-focused mission and its commercial trajectory.12

Founding and the Capped-Profit Pivot

OpenAI was established as a counterweight to the concentration of AI capability in a single corporation, with Musk and Altman as co-chairs, Brockman as chief technology officer, and Ilya Sutskever as chief scientist. Musk departed the board in 2018 amid conflicts of interest with Tesla's own AI work. In 2019 OpenAI created a "capped-profit" subsidiary beneath the nonprofit, took a one-billion-dollar investment from Microsoft, and installed Altman as chief executive, restructuring a nonprofit research mission into an entity capable of raising commercial capital.3 The 2022 release of ChatGPT made OpenAI the central company of the generative-AI period.14

The 2019 capped-profit entity, OpenAI LP, limited early investors to a return of 100 times their money, with lower caps for later backers such as Microsoft; profit above the cap would flow to the controlling nonprofit. Microsoft expanded its commitment to a reported 13 billion dollars by January 2023, receiving a large share of OpenAI's profits up to its cap until repaid. On October 28, 2025 OpenAI completed a further restructuring that retired the capped-profit model entirely: the nonprofit, renamed the OpenAI Foundation, took a roughly 26 percent equity stake valued near 130 billion dollars in a new public-benefit corporation called OpenAI Group PBC, Microsoft took a 27 percent stake worth about 135 billion dollars,5 and employees and investors held the remaining 47 percent, at a 500-billion-dollar valuation. The revised Microsoft agreement moved the power to declare that artificial general intelligence had been reached from OpenAI's board to an independent expert panel jointly established by the two companies.67

The November 2023 Board Crisis

On November 17, 2023, the nonprofit board removed Altman as chief executive, stating that he had not been "consistently candid." The voting directors were Sutskever, Adam D'Angelo, the entrepreneur Tasha McCauley, and Helen Toner of the Georgetown Center for Security and Emerging Technology; McCauley and Toner had ties to the effective-altruism movement, and the board's concern centered on the pace of AI commercialization relative to safety. Brockman quit in protest, roughly 700 employees threatened to resign, and within 48 hours directors discussed merging OpenAI with Anthropic, a step Bloomberg reported Toner had supported and that would have ended OpenAI as an independent entity. Altman was reinstated within days under a reconstituted board chaired by Bret Taylor and including Larry Summers, and Sutskever, McCauley, and Toner left.89

The crisis was widely read as a defeat for the effective-altruism and AI-safety wing of the company by the commercial and accelerationist wing aligned with Altman and Microsoft. Sutskever left OpenAI in 2024, and Musk sued the company and Altman the same year, alleging a betrayal of the original nonprofit mission.810

Musk's suit, filed in 2024 and later amended to name Microsoft as a defendant, alleged that OpenAI had become a "closed-source de facto subsidiary" of Microsoft and accused Altman of "rampant self-dealing." Musk amended the complaint so that any of the roughly 150 billion dollars in damages he sought would go to OpenAI's nonprofit rather than to himself, and sought to remove Altman as chief executive and director. In May 2026 a nine-member advisory jury found that Musk had filed outside the statute of limitations, and Judge Yvonne Gonzalez Rogers dismissed the case.11

The 2026 Recording Device Case

On October 13, 2026 the United States Attorney's Office for the Northern District of California announced that Abdulrahman Alfozan, an OpenAI engineer and a Saudi national, had been arrested on October 9 and charged with using or attempting to use a device to intercept oral communications.12 The Justice Department release states that the complaint alleges he concealed a recording device inside a USB port on the back of a television in an executive conference room at one of OpenAI's San Francisco buildings. OpenAI personnel found the device on September 30, 2026, concealed in toilet paper, and OpenAI determined that it had been installed on or about September 28 and that its contents were encrypted.12

OpenAI said that it had found a "sophisticated" device, which it described as having "advanced encryption, audio monitoring capabilities, and a mechanism for transferring data designed to evade monitoring." The company said that it used its internal models to analyze the device and identified that "a specific employee had planted it a few days earlier," and that it had referred the matter to the FBI.13 On X the company said that "at this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data."13 OpenAI said it fired the employee on October 9, the day of his arrest, after coordinating the timing with the FBI, and that it had restricted his access to systems and spaces beginning September 30.14

OpenAI's spokesperson said the employee started with the company in July 2025 and worked on the product team.13 A LinkedIn profile under his name gives September 2025 as the start of the OpenAI role.15 The company's own GitHub organization lists six pull requests from an account under his handle to the openai-agents-python repository, opened between April 15 and May 7, 2026.16

RAED Ventures, a Riyadh venture firm whose investors include the Saudi Venture Capital Company, Riyadh Valley Company and Jada Fund of Funds (owned by the Public Investment Fund), announced Alfozan as a venture partner on February 8, 2025, when he was a founding engineer at Nash (company). The firm's team page captured on January 23, 2026 describes him as "Engineering Lead at OpenAI, with a focus on AI Agents and the data platform," and the page live on October 13, 2026 carries the same text and an MIT degree; the San Francisco Standard reported that his LinkedIn profile lists a master's degree in data science from the University of California, Berkeley.1317 Prince Khalid bin Bader Al Saud is listed as a venture partner of the same firm; Echelon (company), which he founded, announced a round from East40, RAED Ventures and Avra on October 7, 2026.1718

Gulf Investment and Partnerships

On May 22, 2025 OpenAI announced Stargate UAE, which it described as "the first international deployment of Stargate, OpenAI's AI infrastructure platform," with "a 1GW Stargate UAE cluster in Abu Dhabi with 200MW expected to go live in 2026." The announcement names G42, Oracle, NVIDIA, Cisco and SoftBank as partners in an agreement that it says "was developed in close coordination with the U.S. government."19 The project brings together G42, OpenAI, Oracle, NVIDIA, Cisco and SoftBank under the name Stargate (AI infrastructure) in Abu Dhabi.20 In October 2025 a G42 spokesperson said that "work is progressing rapidly on the first 200 megawatts of capacity, which is expected to be completed by 2026."20

On May 13, 2025 Sam Altman attended the Saudi-U.S. Investment Forum in Riyadh during the visit of President Donald Trump to Saudi Arabia.21 A Reuters photograph of the forum carries the caption "OpenAI CEO Sam Altman looks on as he visits Riyadh with U.S. President Donald Trump and Tesla CEO Elon Musk (both not pictured), in Riyadh, Saudi Arabia, May 13, 2025."21

In October 2025 MGX, an Abu Dhabi investment firm, took part in a secondary sale of about 6.6 billion dollars in OpenAI shares, a transaction that set the company's valuation at around 500 billion dollars. The sale allowed employees and early investors to sell part of their holdings. Other participants named in Reuters reporting were Thrive Capital, SoftBank, Dragoneer Investment Group and T. Rowe Price.22

In March 2026 Bloomberg reported that the company had said Andreessen Horowitz, MGX, D.E. Shaw Ventures, TPG and T. Rowe Price "will co-lead the financing, which is set to close next week." Coatue Management, Thrive Capital and Altimeter Capital were also named as set to participate in the roughly 10 billion dollar financing.23

Government and National Security Engagements

On June 13, 2024 OpenAI added Paul Nakasone to its board. TechCrunch reported that the former National Security Agency head "will also sit on the board's 'security and safety' subcommittee."24

On December 4, 2024 Anduril Industries announced a partnership with OpenAI to apply OpenAI models to counter-drone systems. Anduril said the models "will help protect U.S. and allied military personnel and ensure mission success."25

On January 30, 2025 Los Alamos announced that it had entered a partnership with OpenAI to install its o-series reasoning models on the Venado supercomputer, which was to be moved to a secure, classified network as a shared resource for Los Alamos, Lawrence Livermore National Laboratory and Sandia National Laboratories.26 On August 28, 2025 the laboratory reported that Venado was running OpenAI's o3 reasoning model.27

On June 16, 2025 the Chief Digital and Artificial Intelligence Office of the Department of Defense made an award of 200 million dollars, a prototype other transaction agreement, to OpenAI Public Sector, with an estimated completion date of July 2026. The Department of Defense contracts announcement described the work as meant to "develop prototype frontier AI capabilities to address critical national security challenges in both warfighting and enterprise domains." OpenAI's own description of the work centered on administrative operations, with the aim to "identify and prototype how frontier AI can transform its administrative operations."28

On August 6, 2025 the General Services Administration announced a OneGov partnership with OpenAI under which federal agencies could purchase ChatGPT Enterprise at "$1 per agency pricing" for one year.29

The Justice Department release of October 13, 2026 states that the complaint describes meetings in the executive conference room as involving "nonpublic information about Open AI's technical progress and strategic decisions, capabilities, priorities, government engagements, and commercial relationships," and that the prosecution is in support of NSPM-11, "designed to support American superiority in super intelligence technology."30

Personnel Security and Insider Threat

The New York Times reported on July 4, 2024, citing two people familiar with the incident, that a hacker had accessed an online forum where OpenAI employees discussed the company's latest technologies. Executives told employees and the board at an all-hands meeting in April 2023, and did not inform federal law enforcement agencies, because they "didn't consider the incident a national security threat" and believed the hacker was a private individual with no known connection to a foreign government.31 TechCrunch wrote that unnamed company sources told the Times the hacker "only got access to an employee discussion forum."32

In a June 4, 2024 podcast interview, Leopold Aschenbrenner said that he had written an internal memo on OpenAI's security which he thought was "egregiously insufficient to protect against the theft of model weights or key algorithmic secrets from foreign actors." He said that "a major security incident occurred" a few weeks after he circulated it, that he then shared the memo with board members, and that he received "an official HR warning for sharing the memo with the board." He said that a human resources representative "told me it was racist to worry about CCP espionage and that it was unconstructive." A footnote on the transcript at the words "major security incident" reads: "The previous linked item here was not the major security incident Leopold was referring to."33 TechCrunch reported that Aschenbrenner had hinted at the forum incident in the podcast, calling it a "major security incident."32

On August 4, 2026 the Justice Department's Civil Rights Division announced a combined 3.2 million dollar settlement with OpenAI OpCo LLC and its subsidiary Statsig Inc. over allegations that the companies violated the Immigration and Nationality Act by discriminating against U.S. workers in favor of workers with temporary employment visas during recruiting under the Permanent Labor Certification (PERM) process. The department's investigation found that OpenAI did not advertise PERM positions on its external job website, required paper applications by mail for those positions, and advertised positions on the radio late at night. There were fewer than ten PERM positions at issue. OpenAI agreed to pay 1.2 million dollars in civil penalties and to establish a 2 million dollar back-pay fund, to post PERM positions on its public career website, accept electronic applications, train its personnel, revise its employment policies, and be subject to departmental monitoring and reporting requirements. Reuters reported that the company denied wrongdoing in the settlement agreement.34

In October 2025 OpenAI reported that it had banned accounts apparently tied to Chinese government entities that sought assistance designing social media monitoring tools and profiling dissidents. The same report describes alleged China-linked groups that crafted phishing content and debugged malware targeting Taiwan's semiconductor sector.35

In November 2025 Anthropic reported that a threat actor it assessed with high confidence was "a Chinese state-sponsored group" had manipulated its Claude Code tool into attempting infiltration of "roughly thirty global targets," with success in a small number of cases.36

Footnotes

  1. "OpenAI launches as a nonprofit with a billion-dollar pledge," December 11, 2015; founding donors including Musk, Altman, Brockman, Reid Hoffman, Jessica Livingston, Peter Thiel, AWS, and Infosys. See IEEE Spectrum, December 2015. https://spectrum.ieee.org/billiondollar-nonprofit-ai-research-lab-to-open-in-san-francisco ↩︎

  2. "OpenAI says Musk only ever contributed $45 million," TechCrunch, March 5, 2024, on actual versus pledged funding. https://techcrunch.com/2024/03/05/openai-response-elon-musk-lawsuit/ ↩︎

  3. "Microsoft invests in and partners with OpenAI to support us building beneficial AGI," OpenAI, July 22, 2019, stating that "Microsoft is investing $1 billion in OpenAI" and that "Microsoft will become our exclusive cloud provider." https://openai.com/index/microsoft-invests-in-and-partners-with-openai ↩︎

  4. "The messy, secretive reality behind OpenAI's bid to save the world," MIT Technology Review, February 17, 2020. https://www.technologyreview.com/2020/02/17/844721/ai-openai-moonshot-elon-musk-sam-altman-greg-brockman-messy-secretive-reality/ ↩︎

  5. "The next chapter of the Microsoft-OpenAI partnership," OpenAI, October 28, 2025, stating that Microsoft's investment was "valued at approximately $135 billion, representing roughly 27 percent on an as-converted diluted basis." https://openai.com/index/next-chapter-of-microsoft-openai-partnership/ ↩︎

  6. "Microsoft's $13 billion bet on OpenAI carries huge potential along with plenty of uncertainty," CNBC, April 8, 2023, on the 100x return cap and Microsoft's expanded investment. https://www.cnbc.com/2023/04/08/microsofts-complex-bet-on-openai-brings-potential-and-uncertainty.html ↩︎

  7. "OpenAI completes restructure, solidifying Microsoft as a major shareholder," CNBC, October 28, 2025, on the OpenAI Foundation, OpenAI Group PBC, the 27 percent Microsoft stake, and the AGI declaration panel. https://www.cnbc.com/2025/10/28/open-ai-for-profit-microsoft.html ↩︎

  8. "Effective Altruism's Role in the OpenAI Chaos, Explained," Bloomberg, November 22, 2023. https://www.bloomberg.com/news/articles/2023-11-22/effective-altruism-s-role-in-the-openai-chaos-explained ↩︎

  9. "Inside the Deposition That Showed How OpenAI Nearly Destroyed Itself," Decrypt, on the Sutskever deposition and the discussed Anthropic merger. https://decrypt.co/347349/inside-deposition-showed-openai-nearly-destroyed-itself ↩︎

  10. "OpenAI, Sam Altman and Failed Governance," AKF Partners, postmortem of the November 2023 board crisis. https://akfpartners.com/growth-blog/openai-sam-altman-and-failed-governance-a-postmortem-of-the-openai-governance-soap-opera ↩︎

  11. "Jury dismisses all claims in Elon Musk's lawsuit against OpenAI CEO Sam Altman," NPR, May 18, 2026; "Musk's amended lawsuit against OpenAI names Microsoft as defendant," TechCrunch, November 14, 2024. https://www.npr.org/2026/05/18/nx-s1-5822366/musk-altman-openai-jury-verdict-claims-dismissed ↩︎

  12. U.S. Department of Justice, Office of Public Affairs, "Saudi National Arrested for Installing Surreptitious Audio Recording Device at OpenAI," October 13, 2026, press release 26-1173. https://www.justice.gov/opa/pr/saudi-national-arrested-installing-surreptitious-audio-recording-device-openai ↩︎

  13. TechCrunch, "FBI arrests OpenAI employee accused of planting hidden recording device in conference room," October 13, 2026. https://techcrunch.com/2026/10/13/fbi-arrests-openai-employee-accused-of-planting-hidden-recording-device-in-conference-room/ ↩︎

  14. "OpenAI engineer arrested in bugging of executive conference room," San Francisco Standard, October 13, 2026. https://sfstandard.com/2026/10/13/openai-engineer-arrested-alleged-bugging-executive-conference-room/ ↩︎

  15. LinkedIn, public profile of Abdulrahman Alfozan. https://www.linkedin.com/in/alfozan ↩︎

  16. GitHub, pull requests by the author alfozan in openai/openai-agents-python, queried October 13, 2026. https://api.github.com/search/issues?q=repo:openai/openai-agents-python+author:alfozan ↩︎

  17. RAED Ventures, "About," English about page, Wayback capture of January 23, 2026 (https://web.archive.org/web/20260123220439/https://raed.vc/about/) and live on October 13, 2026 (https://raed.vc/about/); announcement dated February 8, 2025 (https://raed.vc/news/raed-ventures-welcomes-new-venture-partner-abdulrahman-alfozan/). Listing as venture partner first appears in the homepage capture of August 21, 2025 (https://web.archive.org/web/20250821032650/https://raed.vc/). ↩︎

  18. "Saudi AI startup Echelon secures backing from East40, RAED Ventures and Avra," Wamda, October 7, 2026. https://www.wamda.com/ar/2026/10/saudi-ai-startup-echelon-secures-backing-east40-raed-ventures-avra ↩︎

  19. OpenAI, "Introducing Stargate UAE," May 22, 2025 (Wayback capture of May 22, 2025: https://web.archive.org/web/20250522id_/https://openai.com/index/introducing-stargate-uae/). https://openai.com/index/introducing-stargate-uae/ ↩︎

  20. "New update on Stargate UAE: First phase of 200MW to complete by 2026," Gulf News, October 16, 2025. The article states: "Announced in May 2025, the project brings together G42, OpenAI, Oracle, NVIDIA, Cisco, and SoftBank," and quotes a G42 spokesperson: "work is progressing rapidly on the first 200 megawatts of capacity, which is expected to be completed by 2026." https://gulfnews.com/business/markets/stargate-uae-first-phase-of-200mw-to-complete-by-2026-1.500310364 ↩︎

  21. "Photos: Tech CEOs mingle with Trump and Saudi crown prince at investment forum in Riyadh," NBC (NBC San Diego), May 13, 2025. The article names "OpenAI CEO Sam Altman" among the executives present; the Reuters photograph caption reads "OpenAI CEO Sam Altman looks on as he visits Riyadh with U.S. President Donald Trump and Tesla CEO Elon Musk (both not pictured), in Riyadh, Saudi Arabia, May 13, 2025." https://www.nbcsandiego.com/news/business/money-report/photos-tech-ceos-mingle-with-trump-and-saudi-crown-prince-at-investment-forum-in-riyadh/3824637 ↩︎

  22. Reuters report on the secondary sale, as reproduced by Wamda, October 8, 2025: "MGX has participated in a $6.6 billion secondary share sale in OpenAI," at a valuation of "around $500 billion." Named participants: Thrive Capital, SoftBank, Dragoneer, T. Rowe Price and MGX. https://www.wamda.com/en/2025/10/mgx-joins-6-6-billion-openai-share-sale-valuing-chatgpt-maker-500-billion ↩︎

  23. Bloomberg, "OpenAI Set to Raise $10 Billion From MGX, Coatue, Thrive," March 24, 2026, as reproduced by Bloomberg Government: "Andreessen Horowitz, Abu Dhabi's MGX, D.E. Shaw Ventures, TPG and T. Rowe Price will co-lead the financing, which is set to close next week, the company told Bloomberg Tuesday. Coatue Management, Thrive Capital and Altimeter Capital are also set to participate." https://news.bgov.com/california-brief/openai-set-to-raise-10-billion-from-mgx-coatue-thrive-2 ↩︎

  24. Coldewey, Devin, "Former NSA head joins OpenAI board and safety committee," TechCrunch, June 13, 2024: "Former head of the National Security Agency, retired Gen. Paul Nakasone, will join OpenAI's board of directors, the AI company announced Thursday afternoon. He will also sit on the board's 'security and safety' subcommittee." https://techcrunch.com/2024/06/13/former-nsa-head-joins-openai-board-and-safety-committee ↩︎

  25. Defense News, "Anduril, OpenAI join to boost counter-drone tech for US bases, troops," December 5, 2024: "Anduril and OpenAI will partner to improve the military's counter-drone capabilities." https://www.defensenews.com/pentagon/2024/12/05/anduril-openai-join-to-boost-counter-drone-tech-for-us-bases-troops/ ↩︎

  26. Los Alamos National Laboratory, "Los Alamos Lab partners with OpenAI to boost national security," January 30, 2025: the Venado machine "will be moved to a secure, classified network where it will be a shared resource for researchers from Los Alamos, Lawrence Livermore, and Sandia national labs." https://www.lanl.gov/media/news/0130-open-ai ↩︎

  27. Los Alamos National Laboratory, "Lab launches frontier AI models on the Venado supercomputer," August 28, 2025: "the Venado supercomputer at Los Alamos National Laboratory is now running OpenAI's latest o3 reasoning model to accelerate national security research." https://www.lanl.gov/media/news/0828-venado-ai-models ↩︎

  28. DefenseScoop, June 17, 2025: the Department of Defense contracts list announced on June 16, 2025 that the CDAO "had awarded a $200 million prototype other transaction agreement to OpenAI Public Sector" to "develop prototype frontier AI capabilities to address critical national security challenges in both warfighting and enterprise domains"; "The estimated completion date for the work is July 2026"; OpenAI's phrase "identify and prototype how frontier AI can transform its administrative operations." https://defensescoop.com/2025/06/17/pentagon-openai-frontier-ai-projects-cdao/ ↩︎

  29. General Services Administration, "GSA Announces New Partnership with OpenAI, Delivering Deep Discount to ChatGPT Gov-Wide Through MAS," August 6, 2025: "$1 per agency pricing for ChatGPT." https://www.gsa.gov/about-us/newsroom/news-releases/gsa-announces-new-partnership-with-openai-delivering-deep-discount-to-chatgpt-08062025 ↩︎

  30. U.S. Department of Justice, Office of Public Affairs, "Saudi National Arrested for Installing Surreptitious Audio Recording Device at OpenAI," press release 26-1173, October 13, 2026. https://www.justice.gov/opa/pr/saudi-national-arrested-installing-surreptitious-audio-recording-device-openai ↩︎

  31. "Hacker accessed OpenAI's internal AI details in 2023 breach: report," Fox Business, July 7, 2024, summarizing the New York Times report of July 4, 2024: "OpenAI executives told employees and the company's board about the breach at an all-hands meeting in April 2023." https://www-ak-ms.foxbusiness.com/technology/hacker-accessed-openais-internal-ai-details-2023-breach-report ↩︎

  32. Coldewey, Devin, "OpenAI breach is a reminder that AI companies are treasure troves for hackers," TechCrunch, July 5, 2024: "The New York Times reported the hack in more detail after former OpenAI employee Leopold Aschenbrenner hinted at it recently in a podcast." https://techcrunch.com/2024/07/05/openai-breach-is-a-reminder-that-ai-companies-are-treasure-troves-for-hackers/ ↩︎

  33. Aschenbrenner, Leopold, interview with Dwarkesh Patel, June 4, 2024, transcript at approximately 02:33: "egregiously insufficient to protect against the theft of model weights or key algorithmic secrets from foreign actors"; "A few weeks later, a major security incident occurred"; "I got an official HR warning for sharing the memo with the board"; "The HR person told me it was racist to worry about CCP espionage and that it was unconstructive." https://www.dwarkesh.com/p/leopold-aschenbrenner ↩︎

  34. U.S. Department of Justice, Civil Rights Division, press release of August 4, 2026 (https://www.justice.gov/opa/pr/civil-rights-division-secures-settlement-openai-discriminating-against-us-workers); Reuters report datelined August 4, 2026, carried by The Star (Malaysia): "The company denied wrongdoing in the settlement agreement." https://www.thestar.com.my/tech/tech-news/2026/08/05/openai-pays-32-million-in-us-probe-over-hiring-foreign-workers ↩︎

  35. SiliconANGLE, "OpenAI details expanding efforts to disrupt malicious use of AI in new report," October 7, 2025 (summarizing the OpenAI threat report): alleged China-linked groups "were found to be crafting phishing content and debugging malware targeting Taiwan's semiconductor sector, U.S. academia and political groups"; users "sought assistance designing social media monitoring tools, profiling dissidents and generating propaganda proposals." https://siliconangle.com/2025/10/07/openai-details-expanding-efforts-disrupt-malicious-use-ai-new-report/ ↩︎

  36. Anthropic, threat intelligence report published November 13, 2025, stating that the threat actor was "a Chinese state-sponsored group" that targeted "roughly thirty global targets." https://www.anthropic.com/news/disrupting-AI-espionage ↩︎