Former OpenAI director reveals the inside story of Ultraman's recall: The board of directors knew that ChatGPT had been released from X

Tencent Technology

2024-05-29 11:04Posted on the official account of Hebei Tencent News Technology Channel

Highlight the point

• 1

  Helen Toner, a former board member of OpenAI, has revealed the latest details of Ultraman's ouster.

• 2

  Tonna claimed that Altman lied to the board "many times" and "withheld information."

• 3

  Tonna said Altman didn't even tell board members that ChatGPT had been released and that they found it on Twitter.

Tencent Technology News reported on May 29 that according to foreign media reports, Helen Toner, a former member of the board of directors of OpenAI, disclosed the latest inside story about CEO Sam Altman (Sam Altman) who was briefly ousted in November last year.

In an interview that aired Tuesday, Tonna noted that Altman lied to board members on multiple occasions. For example, she said, Altman didn't even disclose ChatGPT's release to the board, and board members learned about it on Twitter.

She further pointed out that Altman has maintained a pattern of behavior of "withholding information" and "distorting facts" for many years, and he has always been selective in disclosing about the situation inside the company.

Read more

Former director of OpenAI wrote an article criticizing Ultraman: he created a "toxic culture of lying" and was suspected of "psychological abuse"

As one of the former board members who voted to remove Altman from his position as CEO, Tonna also accused Altman of hiding the company's ownership structure from the board. "Altman has never disclosed to the board of directors the fact that he owns the OpenAI Startup Fund, although he has always insisted that he is independent as a member of the board and has no financial interests with the company," she said. ”

Tonna believes that Altman's cover-up "seriously damaged our trust in him" and revealed that the board of directors has been seriously discussing whether he needs to be fired since October last year.

Tona, who currently serves as the director of strategy at Georgetown University's Center for Security and Emerging Technologies, noted that Altman's information to the board about OpenAI's security processes was inaccurate and limited. This makes it "almost impossible" for the board to judge whether a company's security measures are adequate or if adjustments are needed.

She said that while there were multiple other examples, the board's final conclusion was: "We can no longer believe what Altman said, and this is a completely unacceptable state for us." ”

Tonna also added that the board can no longer trust Altman's words. She stressed that the board's role is to provide independent oversight of OpenAI, "not just to assist the CEO in raising more money."

According to Toner, during multiple rounds of board discussions last October, several executives used the word "psychological abuse" when describing their experience working with Altman.

Tonna said executives expressed to the board a serious distrust of Altman, who "don't believe is the right person to lead the company toward AGI (Artificial General Intelligence), and they don't believe that Altman has the ability or willingness to make a difference, and therefore don't feel the need to give him feedback or try to fix existing problems." ”

Tonna further noted that when the board realized that Altman had to be replaced, they knew full well that he would do everything in his power to block the decision. Ultraman, she claimed, "started lying to other board members and tried to push me out of the board."

"We acted extremely cautiously and only disclosed the matter to a limited extent, and we hardly informed anyone in advance except the legal team." "That's why on November 17 last year, we suddenly announced the removal of Ultraman. ”

However, Ultraman's ouster did not last long. Less than a week later, the company's board of directors decided to return Altman to the CEO role due to the employee's threat to resign and speculation that Microsoft might poach his team.

Less than two weeks after Altman returned to his role as CEO, Tonna stepped down as a member of OpenAI's board of directors.

After Tonna disclosed the above details, OpenAI did not immediately respond. But current chairman Bret Taylor said in a statement: "We are disappointed that Ms. Tonna continues to revisit these issues. ”

Taylor said that after an independent review, the decision regarding Altman's dismissal "was not based on concerns about product safety, speed of development, OpenAI's financial health, or its statements to investors, customers or business partners." He further noted that "more than 95% of employees" support Altman's reinstatement, saying that the company remains focused on its mission, which is to ensure that AGI can benefit all of humanity.

The reason for Ultraman's dismissal has caused widespread speculation and discussion in Silicon Valley. At the time, the board of directors said only that Altman had failed to maintain "consistent candour" in his communications with the directors. However, in the months that followed, new details emerged about the tension between Altman, the board of directors and some employees.

In an article in The Economist last weekend, Tonna and Tasha McCauley, another former OpenAI director, laid out their thinking, saying that OpenAI is not self-regulating and that governments should intervene to ensure that powerful AI can develop safely and sustainably. (Compiler/Golden Deer)

View original image 567K

• 

•