OpenAI has been exposed: the conspiracy theory of turning from non-profit to for-profit

OpenAI is systematically and premeditating a fundamental shift from "for the well-being of mankind" to "for the profit of investors", CEO Altman has a long-term, well-documented pattern of inconsistent words and actions, manipulation of information and circumvention of oversight, and the deep binding of personal investments to the company's business. (Synopsis: GPT-5 debuts this summer!) Sam Altman reveals OpenAI's next-generation model blueprint, StarGate ambitions, will there be ads? (Background added: OpenAI Shock Bomb!) Acquisition of legendary Apple designer Jony Ive io, Altman: "The most fantastic AI hardware" is on the way) An interactive report of more than 50 pages and more than 10,000 words reveals the evolution of OpenAI from a nonprofit research lab to a for-profit giant. Recently, two nonprofit technology watchdog organizations, the Midas Project and the Tech Oversight Project, released an in-depth investigation called OpenAI Archives. The report, led by Tyler Johnston, executive director of the Midas Project, took nearly a year of public information gathering and a month of intensive writing. The report calls it "the most comprehensive compilation to date of documented concerns about OpenAI in terms of corporate governance practices, leadership integrity, and organizational culture." By combing through a large number of public materials such as company disclosure files, legal proceedings, open letters and media reports, the interactive report of more than 10,000 words found that OpenAI is systematically and premeditating a fundamental shift from "for the well-being of mankind" to "for the profit of investors", with CEO Altman having a long-term, well-documented pattern of inconsistent words and actions, manipulation of information and evasion of supervision, and the deep binding of personal investments to the company's business. OpenAI is at odds when it comes to security and transparency, and its public commitment is grossly disconnected from internal practices. The report is divided into four main themes: restructuring, CEO integrity, transparency and security, and conflict of interest. The extent to which OpenAI executives and board members have directly or indirectly benefited from the company's success is of concern. This includes an analysis of CEO Altman's portfolio involving Retro Biosciences, Helion Energy, Reddit, Stripe, and a number of other companies that have partnerships with OpenAI. Restructuring: An elaborate "betrayal of mission" The report notes that OpenAI is systematically and premeditated to dismantle the core ethical and structural pillars of its founding, and that its actions are seriously at odds with public pronouncements, essentially a fundamental shift from "for the good of humanity" to "for the profit of investors." First, the report reveals the simultaneous disintegration of OpenAI's two core pillars—"profit caps" and "nonprofit oversight." The original "Capped-Profit" model was its core philosophy, which was intended to ensure that the enormous wealth created by AGI was shared with all of humanity and to prevent excessive concentration of wealth. However, this commitment has been gradually hollowed out: from the reduction of profit multiples that ostensibly strengthen the mission, to the secret introduction of a clause of "automatic growth of 20% per year" that makes it functionally useless, to the final plan to completely remove the cap, marking the end of the idea of wealth sharing. At the same time, its oversight mechanisms have been cleverly weakened. OpenAI transformed from an entity fully controlled by a nonprofit to a public benefit corporation in Delaware. The legal obligation has changed from "mission priority" to "balancing shareholder interests and the public interest". The report notes that "there is no precedent in history for shareholders to successfully sue to protect the public interest," making pro bono promises virtually impossible to enforce in legal practice. This suggests that the PBC's commitment to the "public good" may be reduced to a shell in reality, providing a huge legal cover for the pursuit of profit maximization. Image source: openaifiles.org/ website The report further refutes OpenAI's official rhetoric of abandoning its commitment on the grounds that "the industry is highly competitive." Citing the company's early Bylaws and internal emails, the report proves that OpenAI anticipated and was prepared to deal with intense industry competition from its inception. Therefore, using competition as a justification for reneging on promises is an untenable "revisionist history". The real motivation behind this is precisely that investors and company executives believe in its huge profit potential, so it is crucial to remove the cap. CEO Integrity: CEO Behavior Patterns Trigger a Crisis of Trust The report further notes that CEO Altman has a long-standing, well-documented pattern of inconsistent speech and action, manipulation of information and evasion of oversight, and prioritization of personal interests over organizational responsibility. The report cites several examples of Altman openly lying or misleading on major issues. Example: On the issue of employee non-derogatory agreements, Altman publicly claimed to be unaware of the "deprivation of equity of departing employees" clause, but the dossier shows that he explicitly authorized the clause. In Senate testimony under oath, he claimed not to have an equity interest in OpenAI, but later admitted to holding it indirectly through the fund. The fact that he personally owns the OpenAI startup fund has long been hidden from the board. Helen Toner, a former board member, directly accused Altman of obstructing the board's performance by "withholding information, misrepresenting facts, and even outright lying." The report also shows that this pattern of behavior has continued throughout his career: During his time at Loopt, senior employees twice tried to get the board to fire him, citing "deceptive and chaotic" behavior. During his time at Y Combinator, he neglected his duties by focusing on his personal projects, and was eventually "asked away" by founder Paul Graham. The most dramatic manifestation is that after being fired from the OpenAI board, he used his influence to reverse the operation, successfully achieving a "backlash" of the oversight system by "removing his board members and installing his own allies" as a return condition. Operational and Security Risks: Systemic Failure of Security Commitments The report reveals a systemic discrepancy between OpenAI's words and deeds when it comes to security and transparency, with a serious disconnect between its public commitments and internal practices. The company culture shows a tendency to "only fast but not broken", and is systematically weakening, circumventing and even punishing internal safety oversight and dissent in order to pursue business interests and competitive advantages. The report reveals a systematic inconsistency between OpenAI's words and deeds when it comes to security and transparency. The company promised to devote 20% of its computing resources to the "super-aligned" security team, but according to former head Jan Leike, this resource was never allocated. In GPT-4o development, the security team was asked to "fast-track" testing before the product was released, and the company planned the release celebration even before the evaluation began. More seriously, the company threatens departing employees with draconian severance agreements that will cost the company millions of dollars in equity if criticized. Employee Leopold Aschenbrenner was fired for filing a national security risk memo with the board, and the company explicitly informed him that the reason for the dismissal was that he "went above and beyond" to report security issues. The report also pointed out that OpenAI had a serious security incident of hacking and theft of AI technical details in 2023, but it did not report to the company for up to a year.