Sam Altman tailored OpenAI's work processes to his preferences, abandoned the original noble mission, and misled the board of directors.

New Yorker journalists conducted a comprehensive year-and-a-half investigation into Sam Altman's activities and concluded that he frequently lied while serving as CEO of OpenAI.

Investigative reporter Ronan Farrow and New Yorker author Andrew Marantz examined previously unpublished internal memos, 200 pages of documents, and interviewed over 100 individuals.

The primary goal was to understand why Altman was dismissed by board members in November 2023.

“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. The board concluded that Altman lacks these qualities. We wonder: are they right to claim he cannot be trusted?” wrote Farrow.

The authors note that in the fall of 2023, OpenAI's Chief Scientist Ilya Sutskever compiled about 70 pages of memos regarding Altman and his deputy Greg Brockman. One memo begins with the words: “Sam shows a consistent tendency to lie.”

Dario Amodei, who left the company, kept his own personal notes. In one document, he referred to the words of OpenAI's head as “nonsense.”

Those who contributed to Altman's dismissal accused him of deception.

“He builds structures that on paper should limit him in the future. But then, when that moment arrives, Altman discards that mechanism, whatever it may be,” states one document.

How Does Deception Manifest in Practice?

At the end of 2022, Altman assured the board that the functions of the upcoming AI model were approved by the safety committee. Helen Toner requested the relevant documentation and found that the most controversial decisions had not actually been approved.

In 2023, the company was preparing to launch GPT-4 Turbo. At that time, Altman told CTO Mira Murati that the model did not require approval from the safety department, citing the company’s chief legal officer Jason Kwon. However, Kwon “did not understand” where Altman got such an idea.

The article also discusses how OpenAI's leadership considered profiting by pitting global powers against each other, including China and Russia.

The plan was abandoned after several employees expressed their intention to resign.

Another deception involved OpenAI's status as a non-profit organization. It accepted charitable donations, and some employees joined specifically because of the company’s noble mission, even taking pay cuts for it.

However, internal documents show that as early as 2017, the founders had doubts about the non-profit structure. Brockman wrote in his diary:

“I can’t say we are committed to the non-profit model. If we become a B-Corp in three months, that would be a lie.”

In October 2025, OpenAI completed a restructuring that split the company into a commercial corporation and a non-profit fund.

Competition Above All

Some former OpenAI researchers stated that the firm strayed from its original mission of ensuring safety and accelerated an industry-wide “race to the bottom.”

The article details a number of public and internal safety commitments that the company abandoned. Several relevant teams were disbanded.

Recall that in May 2025, during the update of the flagship AI model ChatGPT, OpenAI ignored the concerns of expert testers, making it excessively “sycophantic.”