The End of the Altman Myth and the Collapse of Trust in Global AI Money
~ The True Identity of the 'AI Prophet' Exposed by The OpenAI Files ~
The OpenAI Files have exposed more than just a management scandal.
It is a profoundly structural whistleblower report that uncovers the privatization of power, lack of transparency, and hollowing out of ethics nesting at the core of a tech company once hailed as the 'palanquin' of the 21st century.
Sam Altman—a man who styled himself as the 'prophet of the AI era' and siphoned astronomical sums from global investors by framing AGI (Artificial General Intelligence) as a myth.
However, the legitimacy of OpenAI, which had expanded by relying on that charisma, has been fundamentally overturned by this investigative report.
Microsoft, which has continued to provide massive funding, is finally beginning to perform the 'wisdom of keeping one's distance.' But the problem lies right next to them.
It is SoftBank Group (SBG), which boasts that it will 'remake entire nations with AI' and is still trying to bet on the fantasy known as the 'Stargate Project' with borrowed money.
Now that the AI missile launched by Altman is malfunctioning, it may be none other than SBG itself that takes the full force of the blast.
It was never because they had 'foresight.' They have simply rushed this far on the strength of their own 'deep-seated convictions' alone.
For SBG, this headwind is not bad luck, but an 'inevitability.'
Right now, the vision of the future for global AI infrastructure, with OpenAI at its core, is dissipating.
Now that the governance failures of the 'prophet' Altman have been exposed to the light of day, not only the Stargate Project but the entirety of global AI investment is shifting from 'faith' to 'doubt.'
What is being questioned is not AI technology itself, but the desires of the people swarming around it and the distortions of the system.
The Truth About Sam Altman and OpenAI Exposed by The OpenAI Files
Introduction: Background and Significance of 'The OpenAI Files'
'The OpenAI Files,' released on June 19, 2025, is a comprehensive investigative report by the non-profit organizations Midas Project and Tech Oversight Project [1][2].
This report exhaustively organizes concerns lurking within OpenAI's governance system, leadership structure, and organizational culture, and calls for the necessity of transparency and a responsible management system in AI development [1].
Under CEO Sam Altman, who claims that 'AGI (Artificial General Intelligence) will be realized within a few years,' OpenAI has gathered massive funds from around the world and pushed forward with the development of cutting-edge AI models.
However, in that process, numerous issues have been pointed out, such as the divergence from its non-profit mission, the opacity of its leadership, and a retaliatory culture toward employees.
In this article, based on 'The OpenAI Files,' we will calmly verify Altman's leadership and the reality of OpenAI.
We will delve into the issue from three perspectives: the hollowing out of the non-profit mission, the integrity of management, and the ethical responsibilities of AI development, and discuss the future of AI governance and international regulation.
1. The Truth About Altman and the Y Combinator Chairmanship
First, there is the point regarding the relationship between Altman and the startup accelerator Y Combinator (hereinafter YC).
Altman served as president of YC from 2014 to 2019, but after his departure, it was flagged as an issue that he was listed as Chairman of the Board of YC in documents submitted to the SEC (U.S. Securities and Exchange Commission) [3].
In reality, according to YC, Altman never officially assumed the position of Chairman of the Board, nor was his name listed as a director [3].
This was an error in the biographical description within an Altman-related SPAC (Special Purpose Acquisition Company).
Because it was explicitly stated in the SEC registration documents that 'Altman currently serves as Chairman of the Board of YC' [3], the discrepancy with the facts was exposed.
It has been reported that YC co-founder Paul Graham effectively dismissed Altman in 2019 [9], and given this background, it was unnatural for Altman to call himself 'YC Chairman'.
In short, Altman described himself as YC Chairman in official documents to the SEC, but YC did not approve that position, and he was not the Chairman of the Board [3].
This incident raised doubts about the representation of his career as an executive and cast a stone at the transparency of governance.
2. OpenAI's Profit Cap Rule and the Change to '20% Annual Increase'
OpenAI started as a non-profit organization when it was founded in 2015, and in 2019, it introduced a 'capped profit' model that set a limit on investor returns.
Initially, the pledge was to limit investment returns to a maximum of 100 times, with any excess returned to the non-profit organization for the benefit of all humanity [1][2].
However, the structure changed significantly thereafter.
It was revealed through third-party reporting that OpenAI had quietly relaxed the profit cap for investors and changed the mechanism so that the cap would increase by 20% every year starting in 2025 [2].
This change was not announced by OpenAI itself and was discovered in 2023 through interviews by 'The Information' and 'The Economist' magazine [2].
If the profit cap were to grow by 20% annually, the cap would double in five years, and in the long term, the initially set 100-fold limit would effectively become meaningless [2].
OpenAI had long claimed that it 'limits profit distribution to protect its non-profit mission,' but this cap-raising measure contains a contradiction with that philosophy.
In fact, OpenAI began officially considering the 'abolition of the profit cap' after 2023, and in 2025, it announced a transition to a Public Benefit Corporation (PBC) under the control of the non-profit parent company, which claims to serve public interest, and the complete abolition of profit limits [1][2].
This quiet rule change was a move that prioritized profit distribution to investors and can be called a turning point from the promises made at the time of founding.
3. Altman's 'No Equity' Statement and the Reality of Indirect Holdings
CEO Altman testified at a U.S. Senate hearing in May 2023 (a hearing on AI regulation) that he 'does not hold any equity in OpenAI' [4].
He himself had publicly stated that even after OpenAI transitioned from a non-profit to a 'capped-profit company,' he managed it solely on a salary and held zero equity.
However, subsequent reports revealed that Altman had indirectly held OpenAI shares through venture capital firms such as Sequoia Capital [4].
Specifically, Altman had made an LP (Limited Partner) investment in the fund that Sequoia used to invest in OpenAI in 2021, and although his stake was extremely small, he was in a position to receive economic benefits linked to OpenAI shares [4].
Altman later admitted in an interview, 'I had made a small investment through an old Y Combinator fund, and I also had a very small stake through a Sequoia fund, but I have already sold it' [4].
OpenAI's official website also notes the fact that he had made an indirect investment through a YC fund, explaining that his only economic interest was limited to these early, small-scale investments [4].
In short, although Altman did not hold shares directly, he was, for a time, in a position that could potentially create a conflict of interest with the value of OpenAI's shares.
Regarding this point, an OpenAI spokesperson also acknowledged that 'Altman does not have a direct stake in OpenAI, but he did have a minute exposure through a general Sequoia investment fund' [4].
While his public testimony that 'I do not hold shares' was not literally incorrect, it was incomplete in that it concealed the existence of his indirect holdings, raising questions about governance transparency.
4. Altman's Personal Investments (Reddit Shares, Rain AI, etc.) and Potential Conflicts of Interest
The 'OpenAI Files' also address the point that there may be conflicts of interest between companies Altman is personally involved with and OpenAI's business.
Representative examples are Reddit and Rain AI.
Altman is one of the major shareholders of the social media site Reddit (reportedly holding around a 7% stake), and it is said that he gained hundreds of billions of yen in unrealized gains from Reddit's IPO (Initial Public Offering) in 2023 [5].
In 2024, that same Reddit announced a partnership with OpenAI, beginning the provision of Reddit data to ChatGPT and collaboration in the advertising sector [5].
Regarding this partnership, OpenAI explained that it disclosed 'CEO Altman is a Reddit shareholder,' excluded him from the decision-making process, and proceeded with the approval of an independent board of directors [5].
Even so, the structure in which Altman's personal share value could rise due to OpenAI signing a contract with Reddit was unavoidable.
In fact, Reddit's stock price soared after the partnership announcement, resulting in economic benefits for Altman [5].
Another example is Rain AI.
This is a startup developing neuromorphic (brain-mimicking) chips, in which Altman had personally invested approximately $1 million (over 100 million yen) [6].
In 2019, OpenAI signed a Letter of Intent (LOI) to purchase $51 million (approximately 7 billion yen) worth of AI semiconductor chips to be released by Rain AI in the future [6].
This LOI was signed at a stage when the product had not yet been completed, and it is said that it was used as a credit material for Rain to raise funds, with the claim that 'OpenAI has promised to purchase them in the future' [6].
Needless to say, Altman benefits from the increase in value of Rain, a company he has invested in, raising suspicions that he used OpenAI's resources for personal gain.
Within the company, it has been pointed out that Altman may not have sufficiently informed the board members about this deal with Rain, and it is reported to be one of the backgrounds to his dismissal in November 2023, which will be discussed later [13].
Although the partnership between OpenAI and Reddit proceeded without Altman, the very situation where the top executive's personal investments and company transactions intersect speaks to the difficulty of information disclosure and oversight.
The case of the Rain chip purchase was more blatant and, from a third-party perspective, a clear conflict of interest [13].
These incidents demonstrate the vulnerability of corporate governance at OpenAI and highlight the challenge of how to oversee conflicts of interest among executives.
5. The '7% Equity' Grant to Altman During Restructuring and the $20 Billion Valuation Theory
OpenAI announced plans between 2023 and 2024 to transition from its traditional limited-profit LLC to a Public Benefit Corporation (PBC) through organizational restructuring.
As part of this restructuring, it is reported that a proposal to grant Altman personally 7% of OpenAI's equity is being considered within the company [7].
According to a Bloomberg report, as of September 2024, 'OpenAI is in discussions to grant CEO Altman a 7% stake and convert to a for-profit company,' which would mean Altman would hold ownership for the first time since OpenAI's founding [7].
OpenAI's market valuation was approximately $29 billion in 2023, and is said to be in the estimated $80 billion to $90 billion range in 2024 [5]; if it reaches the hundreds of billions of dollars (e.g., $200 billion, or approximately 30 trillion yen) in the future, 7% would be worth over $14 billion (approximately 2 trillion yen).
The 'worth $20 billion' figure mentioned by users corresponds to 7% of an assumed enterprise value of approximately $300 billion (approximately 40 trillion yen) for OpenAI, which is certainly not an absurd amount.
Altman himself has stated since the founding that his goal is 'to contribute to humanity, not for money,' and he has long held no shares [4] [7], so this massive equity grant could be a turning point in his stance.
OpenAI explains the reason for the restructuring as 'needing to raise more capital and reward investors' [2], but granting a large stake to the top executive is a symbolic event of the departure from non-profit status.
If the board of directors maintained its independence, this type of equity grant to an executive should be a matter for careful deliberation, but in OpenAI's case, it is unclear how much the non-profit board can actually intervene [2].
The possibility that Altman could gain immense personal wealth through the future realization of AGI raises concerns that it may subtly influence decision-making within OpenAI.
OpenAI is facing a difficult situation in how to reconcile its mission to 'prioritize benefits to humanity' with economic incentives.
6. The 2023 Unreported Security Breach Issue and the Dismissal of Whistleblower (Leopold Aschenbrenner)
Despite a serious security breach occurring within OpenAI in 2023, there are suspicions that the management did not report it appropriately.
Leopold Aschenbrenner, who was on OpenAI's security team, wrote a memo within the company after this 'serious security incident,' pointing out that the company's security posture was 'woefully inadequate' and that it was exposed to the risk of theft of 'key algorithmic secrets' by foreign powers [8].
He shared this memo with some OpenAI board members, but as a result, he testified that he was warned by the HR department that 'being concerned about Chinese Communist Party espionage is racist' [8].
Furthermore, Aschenbrenner was fired in April 2024 for what was deemed a 'security breach' after he shared internal brainstorming materials regarding AI safety measures with three external researchers [8].
He himself countered that 'sharing information for feedback purposes is common practice within the company,' and stated that the content labeled as a leak was almost identical to the AI development plans that OpenAI discusses publicly (such as 'solving the alignment problem for super-advanced AI within four years') [8].
OpenAI commented that 'the concerns he raised had nothing to do with his dismissal' [8], but given the circumstances, it is hard to shake the impression that an employee who raised internal safety issues was fired in retaliation.
Furthermore, OpenAI has been criticized for neither disclosing nor reporting this 2023 security incident to the authorities.
According to Aschenbrenner's memo, although the incident increased the risk of confidential information leaking to the outside, the company did not take sufficient measures and was evasive [8].
This issue eventually led to whistleblowing to the U.S. SEC and the Department of Labor, and as described later, developed into a request for an investigation into OpenAI's overall NDA policy [12].
For a company dealing with advanced AI, information security is a lifeline, and it can be said that a major question mark hangs over the judgment of the OpenAI management team that dismissed internal warnings.
7. Stock Clawback Clauses for Departing Employees and Altman's Involvement
It was revealed that around 2023, OpenAI had imposed extremely strict contracts (separation agreements) on departing employees.
These included clauses that effectively forced departing employees to 'not criticize or blow the whistle after leaving,' and if they refused, the company could confiscate all stock options (holdings) they had already acquired [1][10].
When this matter was reported in the media, it drew significant criticism from both inside and outside the company, and CEO Altman apologized on X (formerly Twitter) in May 2024 [10].
Altman's post stated, 'We have never actually stripped anyone of their holdings, and we will not do so in the future for refusing to agree to an NDA. Holdings are holdings, and they should not have been retroactively forfeited. It was a mistake for such a clause to be in the separation agreement document, and I was not aware of it. As a leader, I am very embarrassed, and it is entirely my fault' [10].
However, regarding this explanation by Altman, it has been confirmed that leaked internal documents bear the signatures of both him and the head of legal, and that there existed separation agreements explicitly stating that 'if you do not sign the consent form within 60 days, you will lose the right to retain your vested holdings' [10].
Furthermore, the articles of incorporation of the affiliate that manages OpenAI's holdings also included a clause stating that 'shares can be recovered from former employees for almost any reason,' and these documents were signed by Altman as CEO on April 10, 2023 [10].
In other words, the mechanism for confiscating departing employees' stock, which Altman explained he 'did not know about,' was actually introduced with his own approval [10].
This contradiction further damaged internal trust and cast serious doubt on his integrity.
Later, OpenAI management stated that 'such clauses were a mistake and will be corrected promptly' [10], but the very explanation that they were unaware of (or overlooked) the important contract details of their own company as top management resulted in exposing flaws in the governance system.
8. Repeated Demands for Dismissal at Altman's First Startup, Loopt
Distrust in Altman's leadership has been pointed out since before OpenAI.
It has been revealed that at Loopt (a location-based SNS service), the first startup he founded at age 19, the management team had demanded Altman's removal as CEO from the board of directors multiple times [9].
According to revelations made in 2024 by former OpenAI board member Helen Toner, internal dissatisfaction with 'Altman's deceptive and confusing behavior' at Loopt led the management team to suggest to the board twice that they 'remove him as CEO' [9].
In the end, Altman sold Loopt and gained success as an entrepreneur, but it means that even then, there were words and actions that caused confusion within the organization.
This story was not made public for a long time, but it surfaced in the context of 'actually, he was causing problems at both YC and Loopt' when Altman was temporarily dismissed from OpenAI in November 2023 [9].
In fact, it has been reported that his departure from YC in 2019 was also driven by friction with Graham and a lack of focus on his duties [8].
The episode regarding the demand for his dismissal during the Loopt era demonstrates Altman's tendency to prioritize his own goals, leading to conflicts with those around him.
While he is praised as being 'charismatic and visionary,' it can be inferred that there has long been an underlying risk of him disrupting organizational harmony [9].
When the OpenAI board of directors made the decision to dismiss Altman in 2023, it is possible that his past record at Loopt was one of the factors in their judgment.
Having experienced losing the trust of subordinates and co-founders in the past, how is that being applied to the present? In evaluating Altman's leadership style, the lessons from this Loopt era cannot be ignored.
9. Ilya Sutskever's Concern: 'Not the Person to Press the AGI Button'
Ilya Sutskever, co-founder and chief scientist of OpenAI, had been harboring strong anxiety about CEO Altman's stance since around 2023.
From the summer of that year onward, he frankly appealed to independent board members, saying, 'I don't think Sam is the person who should press the AGI (Artificial General Intelligence) button' [11].
In fact, in the autumn of 2023, Sutskever met individually with three board members at the time (Helen Toner, Tasha McCauley, and Adam D'Angelo) and conveyed serious concerns regarding Altman's leadership [11].
In doing so, he stated, 'Sam is not the right person to be entrusted with the AGI switch' [11].
Sutskever had been a central figure since the founding of OpenAI and was a key player in the development of foundation models such as GPT-4, but at the same time, he had begun to show a strong interest in AI safety [11].
With the emergence of AGI coming into view, it is possible that he felt CEO Altman's decision-making was leaning too heavily toward short-term commercialization or maintaining his own influence.
This statement came to light in 2025, but it is reported that Sutskever had actually pointed out that 'Sam is not the man to press the AGI button' as early as November 2023, immediately after Altman was temporarily dismissed [1].
Although Sutskever himself participated in Altman's dismissal and expressed remorse after facing fierce backlash from employees, it seems the fundamental distrust he felt did not disappear.
Because OpenAI handles technology that will have an immeasurable impact on humanity, the qualities required of its leader are extraordinary.
Sutskever's words are a warning that Altman has not earned enough trust to be entrusted with the 'final switch to stop AI from running out of control' [1].
10. Mira Murati's Concern: 'Discomfort with Sam Leading AGI'
Mira Murati, who was the CTO (Chief Technology Officer) of OpenAI, was also one of those who held strong doubts about CEO Altman.
In the fall of 2023, separately from Sutskever, Murati contacted independent board members and told them, 'I am worried (I feel uncomfortable) about being led to AGI by Sam' [11].
She was in a position to oversee the model safety review process within the company, and she had concerns that Altman would often try to rush releases by skipping safety procedures [11].
For example, it is reported that when Altman said, 'GPT-4 Turbo can be released without a review by the Safety Review Committee (DSB),' a check with the legal department revealed that no such approval had actually been obtained [11].
Because of this chaos, Murati came to feel unbearable anxiety about continuing AGI development under Altman's command.
Her statement to the board members that 'I cannot feel at ease entrusting AGI to Sam' is an expression of that frank sentiment [11].
Although Murati was in charge of key technical strategy under the Altman regime, she simultaneously grew increasingly alarmed by the deterioration of the organizational culture (secrecy and disregard for procedures) and worked with Sutskever to try to rectify the situation internally [11].
As a result, she became interim CEO after Altman's dismissal, but stepped down after a short period due to pressure from mass resignations of employees. She subsequently left OpenAI in 2024.
The 'discomfort' Murati felt was a lack of trust in Altman as the top executive, and the fact that even the most capable executive in the technical field felt this way must be taken seriously.
11. Criticism by the Amodei siblings: Altman's behavior is 'gaslighting' and 'psychological abuse'
OpenAI once had siblings Dario Amodei (former Vice President) and Daniela Amodei (former Research Engineer) on staff, who left the company in the late 2010s to co-found Anthropic.
Although they have not spoken publicly about it directly, testimonies from inside and outside the company reveal that the Amodei siblings also held strong distrust of Altman's management style.
It was sometimes described as, 'Altman's management style is gaslighting (psychological manipulation that confuses people and makes them doubt themselves) and is like psychological abuse' [11][14].
In fact, the evidence submitted to the board by the aforementioned Murati, Sutskever, and others reportedly included examples of Altman repeatedly engaging in words and actions that pitted subordinates against each other and created discord [14].
For example, it has been pointed out that in one instance, Altman told an executive that 'another executive is trying to get you fired,' potentially engaging in behavior that fueled mutual distrust.
Such insidious information manipulation is truly worthy of being called 'gaslighting'.
According to a Washington Post report, Altman had a tendency to make employees 'compete with each other in unhealthy ways' [14], and this was one of the major reasons why the board of directors made the decision to dismiss him.
Multiple people have also testified in internal feedback that 'Altman has a pattern of manipulating and confusing people' [11][14].
Furthermore, former board member Toner revealed that in October 2023, two executives (Murati and Sutskever) went so far as to tell the board that 'Altman's behavior was psychological abuse' [11].
It is extraordinary for such expressions to come from within the company, indicating that trust in Altman had been decisively damaged.
Although the Amodei siblings themselves have not commented publicly, it is speculated that their disappointment in Altman was behind their departure from OpenAI.
In any case, the situation where top management's communication is perceived as 'abusive' is a failure of governance in itself, and it speaks to how chaotic OpenAI's internal culture had become.
12. Negative feedback from five or more other executives and employees
It is reported that not only the two top executives, Sutskever and Murati, but at least five other executives and staff at a level close to Altman within OpenAI had anonymously conveyed similar complaints and warnings to the board of directors [11].
This was not merely a rebellion by a few, but evidence that widespread dissatisfaction had been building up within the organization.
According to an article in The Atlantic at the end of 2023, the three independent directors (Ms. Toner, Ms. McCauley, and Mr. D'Angelo) had received similar feedback from 'at least five people one or two levels below Altman' over the past year or two [11].
The content of all of these was to the effect that 'Altman lacks consistency and integrity, does not share important information,' and 'disrupts the team' [11].
Since its founding in 2015, OpenAI has had many talented researchers and executives, but talent attrition has increased, especially after it transitioned to a for-profit company in 2019.
For example, founding members such as Jack Clark, who transitioned from an anthropologist to an AI ethics researcher, also left in 2020, citing differences in direction and organizational opacity as reasons.
Also, the group of researchers who launched Anthropic (including the Amodei siblings) had a background of opposition to OpenAI's 'commercial priority over safety' stance.
After the dismissal drama in November 2023, it was noted that the majority of OpenAI employees signed a petition demanding Altman's return, but behind the scenes, there were voices saying that 'there was peer pressure' and 'they had no choice but to sign because if he didn't return, the sale of unlisted shares (the employees' economic interests) would be blown away' [14].
Contrary to their outward support, it is highly likely that many employees harbored concerns in their hearts.
The fact that at least five or more executives provided negative feedback to the independent directors tells us that trust in Altman as an organization had been significantly shaken [11].
It can be said that the board of directors' sudden decision to dismiss him in November 2023 was because they could not ignore these long-standing voices.
13. Altman's personal ownership of the OpenAI Startup Fund and concealment from the board
In 2021, OpenAI launched the 'OpenAI Startup Fund,' a venture fund to expand its technology ecosystem.
On the surface, it was considered a 'fund managed and operated by OpenAI,' and it has been investing in startups with funding from Microsoft and others.
However, it was revealed in 2023 that this fund was actually established with Altman as the personal owner [11].
Initially, many of the board members did not know the detailed structure of the fund, and it is said that doubts arose when a director overheard a conversation at a party asking, 'Isn't it inappropriate for the returns of the OpenAI Startup Fund to go to Altman personally rather than to OpenAI shareholders?' [11].
As the board of directors questioned Altman over the following months, it was discovered that he himself was the GP (General Partner) of the fund and was raising and managing money from external LPs [11].
Typically, the GP of a VC fund is structured to receive a certain percentage of operating profits as a success fee (carried interest), and if Altman were in a position to receive this, an incentive would arise to utilize OpenAI's technology and information for fund management.
In fact, this Startup Fund was conducting investment activities while enjoying a privileged position, such as early access to OpenAI's unreleased models [11].
The board of directors repeatedly asked Altman for explanations, but it is reported that he initially gave evasive responses, such as saying, 'I just created a tax-efficient structure and there is no financial gain' [11].
OpenAI also explained that 'Altman has no financial interest in this fund and it was only privately owned to establish it quickly' [11], but the independent directors could not simply believe his claims [11].
This is because if there were truly no gain, he could have simply shown the structure transparently to the board from the beginning, and there would be no reason to hide it.
Ultimately, this distrust is also considered to have led to the dismissal drama.
It is a serious governance issue for an executive to privatize company-related business and not fully disclose it to the board of directors.
In Altman's case, it was a situation that could be taken as 'personally profiting from the OpenAI brand' through the Startup Fund, so it is natural that the board of directors felt a sense of crisis [11].
This case is an episode that symbolizes the internal situation of OpenAI, where a profit-seeking structure was born behind the scenes while upholding non-profit ideals.
14. Altman's Fabrication of Board Statements and Findings from Independent Investigation
As the friction with the board of directors deepened, Altman also engaged in the behavior of conveniently distorting the intentions of other directors.
According to the testimony of former director Helen Toner mentioned above, just before the dismissal turmoil, Altman tried to feed false information to other directors, such as 'another director wants to remove you (Toner) from the board' [11].
In other words, there is a suspicion that he tried to make the directors suspicious of each other and create a hostile relationship.
Toner states that she confirmed the facts by checking with the individuals involved and realized that Altman was lying to break the trust between board members [11].
This is an unforgivable act as a top executive, and it was the deciding factor that made the board of directors feel that 'nothing he says can be trusted anymore' [11].
In fact, the official reason announced for the dismissal resolution in November 2023 was that 'Altman was not consistently candid with the board' [14].
It is inferred that the background to this was the false explanation to the directors as just described.
After Altman's dismissal, the newly invited interim management team conducted an investigation by an independent law firm to verify whether there was any misconduct or error in the board's judgment.
As a result, it was concluded that there was no 'serious misconduct' in Altman's actions that would require immediate dismissal, and he returned to the CEO position within a few days [14].
However, this investigation did not deny the fact that there was clever internal maneuvering as described above.
Rather, looking at the process by which Altman, after his return, wiped out the old board including Toner and built a new system under his complete control, his skill in organizational politics becomes apparent.
It can be said that the risk that information manipulation by management could hollow out the role of independent directors has been exposed.
From a governance perspective, providing false explanations to the board of directors is also a betrayal of shareholders and society.
In the case of OpenAI, although it is a private company, its social influence is significant, and lies to the board of directors lead to lies to the public.
This action by Altman ultimately led to the collapse of the board of directors itself and significantly damaged OpenAI's transparency and credibility.
15. OpenAI's NDA (Non-Disclosure Agreement) Issue: Clauses Waiving Rights to Whistleblower Rewards and Appeals to the SEC
OpenAI imposed stricter non-disclosure and non-disparagement clauses on its employees than typical companies, and among these, suspicions arose that it included provisions forcing employees to waive their rights to whistleblower rewards established by the U.S. Securities and Exchange Commission (SEC) [12].
In July 2024, a group of whistleblowers who were former OpenAI employees sent a letter to the SEC, alleging that OpenAI's employee and severance agreements might infringe upon the rights to whistleblowing protected by federal law [12].
Specifically, it is alleged that the contracts included clauses requiring employees to pledge not to accept rewards (financial incentives for those who cooperate in exposing illegal activities) that they would otherwise receive if they reported violations to government authorities such as the SEC [12].
There have been previous cases where the SEC has fined other companies for such clauses, as they are considered to be in violation of federal law (the Dodd-Frank Act).
In the case of OpenAI as well, the whistleblowers strongly urged that 'the SEC should impose fines on OpenAI for each inappropriate contract' [12].
The SEC has not commented on whether an investigation is underway, but the office of prominent Senator Chuck Grassley was also involved in drafting this letter, viewing OpenAI's pressure on whistleblowers as a problem [12].
This move is also linked to the aforementioned case of Leopold Aschenbrenner, and OpenAI's entire NDA policy is under scrutiny.
OpenAI announced in May that it would establish a 'Safety and Security Committee' to strengthen internal controls [12], but the move to the SEC occurred immediately after that.
According to reports, the contracts in question also contained wording that could be interpreted as 'employees must obtain prior permission from the company when providing information to federal authorities' and 'reporting to the SEC could also be subject to criticism' [12].
If these are true, it is an act that forces employees to keep misconduct internal and prevents cooperation with public regulatory authorities, which is extremely unjust.
The SEC's final decision has not yet been released, but this incident is recognized as a symbol of the authoritarian nature of OpenAI's information control.
Precisely because the potential risks of AI are so great, whistleblowing is an important safety valve, and it must be said that contract practices that discourage it exceed the bounds of social acceptability.
16. The Pretext of Supporting Regulation and Hollowing Out Through Lobbying
Altman has visited countries around the world to preach the 'necessity of AI regulation,' acting outwardly as an active supporter of regulation.
However, various reports have exposed that behind the scenes, OpenAI and Altman himself were engaged in lobbying efforts to hollow out regulatory proposals [13][14].
A prime example is the influence exerted on the EU's comprehensive AI Act.
In June 2023, Time magazine reported, based on documents it obtained, that 'OpenAI proposed amendments to the EU to relax the AI Act behind the scenes, and these were actually reflected in the bill' [14].
Specifically, OpenAI argued that 'general-purpose AI systems' like its own GPT-3 and DALL-E 2 should not be classified as 'high-risk AI' under EU law, attempting to avoid the obligations of transparency and accountability imposed on high-risk systems [14].
This argument was aligned with the lobbying efforts of Microsoft and Google, and as a result, the final version passed by the EU Parliament established a separate category called 'General Purpose AI (Foundation Model),' significantly relaxing the application obligations [14].
Furthermore, in May 2023, Altman stated in London that 'OpenAI would not rule out withdrawing from the EU if things continue this way,' putting pressure on European authorities, only to retract the statement the following day [14].
In the United States as well, although Altman proposed a 'government licensing system for AI developers' at a Senate hearing, he has since strengthened his stance against excessive regulation in subsequent statements [11].
In May 2025, he even stated that 'requiring government permission for open-source AI development is a disastrous policy that would cause the U.S. to lose its technological lead' [11].
In other words, while Altman welcomes regulations that are favorable to him, he strongly opposes regulations that would put OpenAI at a competitive disadvantage.
Some point out that the international AI supervisory body concept he advocates is also aimed at cementing the oligopoly of giant tech companies.
OpenAI has also sent lobbyists to the U.S. Congress and administrative authorities, influencing the wording of AI legislation.
This divergence between 'stated principles and actual intentions' is seen throughout the industry, but OpenAI stands at the forefront of it.
The statement that they 'support regulation' remains a pose for the media, and the reality is that they are working to hollow out regulations [14].
It is pointed out that this duality could undermine trust from policymakers and, in turn, have the counterproductive effect of inviting even stronger calls for regulation.
Comprehensive Analysis and Commentary: The Light and Shadow of Leadership, the Contradiction of Non-profit Status, Transparency and Ethics
What emerges from the verification of the above 16 items is a structure where the charisma and the problems of a leader named Sam Altman are two sides of the same coin.
On one hand, his foresight and ability to take action led OpenAI to the great success of ChatGPT and pushed it to the front runner in the AI development race.
However, on the other hand, in that process, the transformation of organizational philosophy and confusion in internal governance have occurred, leading to serious situations such as the mass departure of executive employees and conflict with the board of directors.
The light part of Altman's leadership lies in the fact that he accelerated innovation through bold fundraising and rapid product commercialization.
He has led the organization by upholding the lofty mission of 'benefit to all of humanity' and publicly stating that he disregards personal financial gain.
However, in reality, it appears that he has been strongly conscious of expanding his own influence and increasing OpenAI's corporate value, often driving the organization using methods that could be described as heavy-handed.
It is possible that the self-centered and arbitrary management style pointed out since his days at Loopt and YC has been repeated in the massive project that is OpenAI.
As multiple testimonies suggest, there are indications that Altman has told subordinates what they wanted to hear to make them comply, and when friction arose, he acted in ways that damaged the other party's credibility [11].
These actions created a chain of distrust within the organization, and it is hard to argue against the perception that this constitutes 'gaslighting' that undermines psychological safety.
Ironically, the leader himself ended up shaking the very unity and trust of the team that is most critical for AI development.
Regarding OpenAI's non-profit nature, a major contradiction also stood out.
At the time of its founding, the organization made a pledge to 'prioritize the benefit of humanity over investor profits' and set strict caps on profit distribution.
However, in the face of market competition and the need for massive amounts of capital, that promise was gradually hollowed out.
The 100x profit cap is effectively being abolished, and the control exerted by the non-profit parent company is becoming a mere formality [2].
The story that Altman himself would hold no equity and remain dedicated has ultimately led to a conclusion involving the granting of a 7% stake [7].
Of course, there are aspects where organizational changes are unavoidable as an entity grows.
However, it cannot be denied that there was a lack of effort to honestly explain the gap between the founding vision and the current state, and to gain the understanding of stakeholders.
The stance of emphasizing to the outside world that 'the mission has not changed' while pushing forward with the transformation from non-profit to for-profit only amplified the sense of opacity.
While such contradictions are a challenge common to the entire AI industry, OpenAI has become a target of criticism precisely because it attracts so much attention.
In terms of transparency and ethics, the lessons from OpenAI are extremely important.
How to handle internal safety concerns and how to balance confidentiality with the public interest are difficult problems that AI development companies will face in the future.
OpenAI uses vast amounts of data to train advanced AI models, and its social impact is immense.
Therefore, it was in a position that should have required more open discussion and stricter internal controls than anyone else.
However, the reality was that secrecy intensified, and it became clear that they were operating under contracts that forbade employees from speaking out and even hindered cooperation with authorities [12].
Under these circumstances, there is no way they could gain trust from the outside.
In fact, starting around the spring of 2023, various media outlets published a series of exposés on OpenAI, alleging that the company was prioritizing commercial interests over safety and engaging in reckless development competition [2].
OpenAI's upper management dismissed these reports and brushed off critics as 'doomers,' but this eventually backfired, leading to an organizational crisis in the worst possible form: the firing of its CEO.
The case of OpenAI serves as a wake-up call regarding how technology companies should be governed.
Even for private companies, when the social impact is immense, corporate governance must consider not only shareholder interests but also the public good.
The dysfunction of the independent board of directors, which was supposed to fulfill this role, was rooted in a breakdown in communication between Altman and the board, as well as a skewed power balance.
The OpenAI board consisted of only four to six members, and aside from Altman, they were external experts; however, their access to management information was limited, and they were isolated in the face of Altman's charisma and the support of the employees.
The decision to fire him in November also appeared abrupt and unrealistic, and it was ultimately overturned by the solidarity between Altman and the employees (as well as the backing of Microsoft).
This outcome demonstrates the reality that 'startup-like speed' and 'cautious governance' are difficult to reconcile.
In an organization led by a visionary with such a strong personality as Altman, it is extremely difficult to maintain an independent oversight function.
Conclusion: Future Governance Challenges and Implications for Regulation
The greatest lesson from 'The OpenAI Files' is that conventional corporate governance is insufficient for organizations dealing with cutting-edge technologies that affect the future of humanity.
Especially for technologies like AGI, which have extreme economic, military, and ethical implications, a new governance model that transcends the boundaries of for-profit and non-profit structures is required.
A 'quasi-public organization' like OpenAI can exert global influence without being subject to regulatory oversight or democratic accountability.
In that sense, the OpenAI case has reached a stage where it should no longer be viewed as a governance issue for a single company, but as a global public policy challenge.
Meanwhile, the United States currently lacks a permanent agency to oversee AI governance, and the regulatory framework remains underdeveloped.
OpenAI has exploited this vacuum by operating under the banner of 'self-regulation' and 'moral mission,' but the limitations of this approach have now been exposed.
If this situation continues, it could solidify a pattern where other companies also 'race toward irresponsible AI development for the sake of profit.'
While the absence of regulation may contribute to accelerating innovation, it also carries the risk of leading to a lack of safety and social consideration.
The series of problems at OpenAI can be positioned as a typical example of what happens in the 'vacuum zone' before legal systems catch up.
That is precisely why governments in the U.S., EU, Japan, and elsewhere must now institutionally incorporate elements such as 'accountability,' 'transparency,' 'oversight mechanisms,' and 'involvement of public interest representatives' into AI governance.
Rather than leaving it entirely to for-profit companies, at least the following frameworks are required:
・Legislation of transparency standards (disclosure of model design philosophy, data sources, etc.)
・Mandatory safety audits and reviews by external committees
・Governance participation by representatives of stakeholders other than shareholders (citizens, researchers, users)
・Establishment of a global AI oversight system through international cooperation
Regarding the last point in particular, Altman himself has proposed an 'international AI agency,' but it is crucial to design an effective system so that it is not used to strengthen the hegemony of GAFA or OpenAI [11].
Balancing regulation and innovation is a difficult challenge, but looking at the case of OpenAI, the reality has become clear that sustainable and fair technological development cannot be achieved by relying solely on 'self-discipline'.
From now on, we are entering an era where the human capacity for institutional design—how to 'govern' and 'integrate' AI into society—will be tested more than the evolution of AI itself.
The glory and chaos of OpenAI warn us that the future will become uncontrollable unless technology, organization, ethics, and social design are linked together.
References
[1] TechCrunch – The ‘OpenAI Files’ push for oversight in the race to AGI (June 18, 2025)
[2] The OpenAI Files (Midas Project & Tech Oversight Project) – Restructuring Concerns (June 2025)
[3] San Francisco Business Times – Sam Altman is not on YC’s board. So why claim to be its chair? (Apr. 15, 2024)
https://www.bizjournals.com/sanfrancisco/inno/stories/news/2024/04/15/sam-altman-y-combinator-board-chair.html
[4] TechCrunch – Sam Altman once owned some equity in OpenAI through Sequoia (Dec. 19, 2024)
[5] Business Insider – Reddit announces another big data-sharing AI deal — this time with OpenAI (May 16, 2024)
[6] Business Insider – OpenAI is paying out $51 million to a chip startup that Sam Altman personally invested in, report says (Dec. 4, 2023)
[7] Bloomberg – OpenAI Is Said to Weigh 7% Stake for Sam Altman in Switch to For-Profit (Sep. 25, 2024)
https://www.bloomberg.com/news/articles/2024-09-25/openai-cto-mira-murati-says-she-will-leave-the-company
[8] Business Insider – Ex-OpenAI employee speaks out about why he was fired: ‘I ruffled some feathers’ (June 7, 2024)
[9] Benzinga – Former OpenAI Board Member... says CEO was fired from Y Combinator and another startup for ‘deceptive and chaotic behavior’ (May 29, 2024)
[10] Vox – Leaked OpenAI documents reveal aggressive tactics toward former employees (May 22, 2024)
[11] The Atlantic – ‘We’re Definitely Going to Build a Bunker Before We Release AGI’ – The true story behind the chaos at OpenAI (May 15, 2025)
[12] Reuters – OpenAI whistleblowers ask SEC to investigate alleged restrictive non-disclosure agreements (July 15, 2024)
https://www.reuters.com/technology/openai-whistleblowers-ask-sec-investigate-restrictive-non-disclosure-agreements-2024-07-13/
[13] Reddit (User comment by norcalnatv) – OpenAI agreed to buy $51M of AI chips from a startup (Rain) backed by Sam Altman (Nov. 2023)
https://www.reddit.com/r/hardware/comments/18a8a90/openai_agreed_to_buy_51_million_of_ai_chips_from/
[14] TIME – Exclusive: OpenAI Lobbied the E.U. to Water Down AI Regulation (June 20, 2023)
Rintaro Takechi