A sprawling investigative report has sent shockwaves through Silicon Valley, raising serious questions about OpenAI and the stewardship of its mercurial CEO, Sam Altman.
The New Yorker on Monday published a damning portrait of Altman, drawing from more than 100 interviews and hundreds of pages of internal documents, including the infamous “Ilya Memos.” The investigation alleges that the leader of the world’s most influential artificial intelligence (AI) firm is a “sociopath” whose leadership is defined by manipulation, chronic dishonesty, and a reckless disregard for the existential safety of artificial intelligence.
The report highlights a jarring paradox in Altman’s character. Former board members and colleagues describe him as possessing two traits rarely seen in tandem: an intense, almost pathological desire to be liked in every interaction, coupled with what one insider called a “sociopathic lack of concern for the consequences” of the deceptions required to maintain that image.
The 18-month investigation centers on a trove of confidential Slack messages and HR documents compiled by OpenAI co-founder and former Chief Scientist Ilya Sutskever. These documents, intended for the board, outline a consistent pattern of behavior that Sutskever summarized as “lying.”
According to the report, Altman’s alleged dishonesty was not merely incidental, but a strategic tool used to manage both peers and rivals.
The documents detail incidents where Altman reportedly offered the same high-level job to two different people simultaneously, gave conflicting accounts of company events, and misrepresented safety protocols to the board. Dario Amodei, the former head of research who left to found rival firm Anthropic, echoed these sentiments in his own private notes, saying flatly, “The problem with OpenAI is Sam himself.”
The most significant allegations involve Altman’s treatment of AI safety, the pillar upon which OpenAI was founded. While Altman has publicly lobbied the U.S. Senate for strict AI regulation, the report alleges he was simultaneously working behind the scenes to dilute European AI laws.
Internally, the discrepancy was even more pronounced. OpenAI had originally promised that its superalignment team tasked with ensuring AI remains beneficial to humanity would receive 20% of the company’s total computing power. Insiders reveal that the actual allocation was 1% to 2%. This marginalization led to the high-profile departure of safety lead Jan Leike, who lamented that “safety culture and processes have taken a backseat to shiny products.”
The report also sheds new light on the chaotic November 2023 firing and subsequent reinstatement of Altman. While the board initially cited a “lack of transparency,” the New Yorker suggests the board was essentially “checkmated” by a counter-campaign organized by Altman and investors like Microsoft Corp., which had $13 billion at stake.
A subsequent investigation by the law firm WilmerHale resulted in no written report, a move six insiders claim was designed specifically to suppress transparency regarding the integrity allegations. As OpenAI moves toward a potential $1 trillion valuation and transitions into a fully for-profit entity, the safeguards designed to keep the company accountable, its nonprofit charter and independent board appear to have been systematically dismantled.
Ultimately, the investigation suggests that Altman’s primary motivator is not the philanthropic safe AGI he pitches to the public, but a pursuit of influence.
As one former employee recalled, Altman once admitted that he is not driven by money, but by power.
OpenAI issued the following statement on the New Yorker story: “Much of the piece revisits previously reported events through anonymous claims and selective anecdotes sourced from people with clear agendas. “
