An investigative report reveals serious allegations against Sam Altman, highlighting concerns over his leadership and integrity at OpenAI.
An in-depth investigation by The New Yorker reveals troubling aspects of Sam Altman's leadership at OpenAI, suggesting a departure from the organization's founding principles. The report, authored by Ronan Farrow and Andrew Marantz, is based on 200 pages of internal documents and interviews with over 100 sources.
The investigation primarily aims to elucidate the reasons behind Altman's abrupt dismissal by the board in November 2023. According to the report, board members concluded that Altman lacked the integrity necessary for a role that carries significant responsibility in the realm of artificial intelligence.
“OpenAI was founded on the premise that artificial intelligence could become the most dangerous invention in human history, so the CEO must be a person of extraordinary integrity,” Farrow noted, questioning whether Altman can be trusted.
Internal memos reveal serious concerns about Altman's honesty, with one memo asserting, "Sam shows a consistent tendency to lie." Former OpenAI Chief Scientist Ilya Sutskever compiled around 70 pages of notes detailing issues with Altman and his deputy, Greg Brockman.
Critics of Altman accused him of misleading the board regarding safety approvals for AI models. For instance, in late 2022, he claimed that the safety committee had approved upcoming AI functions, which was later found to be untrue. In 2023, as the company prepared to launch GPT-4 Turbo, Altman told the CTO that the model did not require safety department approval, a claim contradicted by the company’s legal counsel.
The report also suggests that OpenAI's leadership considered strategies that could manipulate global relations, including leveraging tensions between major powers like China and Russia, but abandoned these plans after employee pushback.
Furthermore, the investigation sheds light on the organization's original non-profit mission, which many employees joined under the belief that they were contributing to a noble cause. However, internal documents indicate that doubts about this non-profit model surfaced as early as 2017, with Brockman expressing in his diary the possibility of transitioning to a B-Corp structure.
By October 2025, OpenAI had restructured, separating into a commercial entity and a non-profit fund, further distancing itself from its foundational ideals. Some former researchers expressed that the company had deviated from its commitment to safety and contributed to a detrimental competitive rush in the AI industry, abandoning several safety initiatives in the process.
