A former engineer at ChatGPT-parent OpenAI, Daniel Kokotajlo, has now revealed why he stepped down. He also expressed a lack of faith in the company’s approach to artificial general intelligence and raised concerns over the silencing of researchers.
On Tuesday, Kokotajlo took to X, formerly Twitter, and in a thread of 15 posts, expressed his dissatisfaction with the Sam Altman-led company.
In April, I resigned from OpenAI after losing confidence that the company would behave responsibly in its attempt to build artificial general intelligence — ‘AI systems that are generally smarter than humans,’ he said.
Kokotajlo said he joined OpenAI with the expectation that the organization would significantly increase its investment in safety research as its systems advanced. However, OpenAI never made this shift. People started resigning when they realized this. I was not the first or last to do so.
He went on to say that OpenAI asked him to sign a nondisparagement clause upon his departure, which would have barred him from criticizing the company. It was clear from the paperwork and my communications with OpenAI that I would lose my vested equity in 60 days if I refused to sign.
But, he declined to sign. Kokotajlo then raised concerns about the lack of oversight over the technology developed by labs like OpenAI. He stated, Silencing researchers and making them afraid of retaliation is dangerous when we are currently some of the only people in a position to warn the public.
He also applauded OpenAI’s decision to change these policies. Kokotajlo concluded his statement by advocating for greater transparency from AI labs and urged his former colleagues to continue pushing for transparency from within the company.
Kokotajlo’s resignation comes amidst a series of departures from OpenAI including Ilya Sutskever and Jan Leike, leaders of the superalignment team AI safety team.
Sutskever was originally hired by Elon Musk when the company was founded in 2015. He played a crucial role in removing Altman as CEO in November last year. However, following Altman’s reinstatement, Sutskever announced that he would not return to his previous role as a board member.
Meanwhile, in May it was reported that several OpenAI employees raised concerns after a new report revealed that some former staffers were required to sign non-disparagement agreements linked to their company shares.
OpenAI later reversed its stance and said, Regardless of whether you executed the agreement, we write to notify you that OpenAI has not canceled, and will not cancel any vested units. This adjustment also pertains to current employees.
