Former OpenAI Employees Tear into Sam Altman’s Character During Testimony for Elon Musk

 

“Sam Altman faced damaging testimony from former OpenAI colleagues during the second week of Elon Musk’s trial against the AI company and its CEO.

 

Business Insider reports that the trial between Elon Musk and OpenAI concluded its second week with several witnesses providing critical testimony about Sam Altman’s leadership and the company’s commitment to AI safety.

 

Musk alleges that Altman and OpenAI President Greg Brockman violated the organization’s nonprofit mission by forming a partnership with Microsoft, essentially looting the charity they founded together in 2015.

 

Three key witnesses took the stand this week, each raising concerns about different aspects of Altman’s management of OpenAI. Their testimony focused on AI safety practices, honesty in leadership, and adherence to the organization’s original nonprofit mission.

 

Rosie Campbell, a former AI safety researcher who worked at OpenAI from 2021 to 2024, testified about what she perceived as the company’s declining commitment to safety research. When Campbell began her tenure at OpenAI, the organization maintained two dedicated teams focused on long-term AI safety. One team worked to ensure AI alignment with human values, while Campbell’s team concentrated on preparing the world for superhuman artificial intelligence.

 

According to Campbell’s testimony, OpenAI gradually shifted toward a more product-focused approach. Both long-term AI safety teams were eventually eliminated, and approximately half of Campbell’s team chose to leave the company rather than accept different positions within OpenAI.

 

Campbell also addressed her role in the letter calling for Altman’s reinstatement after the OpenAI board initially ousted him as CEO. She explained that she signed the letter not out of support for Altman, but from concern that OpenAI employees might otherwise end up working at Microsoft, which she believed would be even less committed to AI safety than OpenAI. “It was my understanding at the time that the best way for OpenAI to not disintegrate and fall about would be for Sam to return,” Campbell stated.

 

Interestingly, Campbell also testified that she believed xAI, Musk’s own artificial intelligence company, likely had an inferior approach to safety compared to OpenAI.

 

Tasha McCauley, a former OpenAI board member who participated in the initial decision to remove Altman as CEO, provided testimony through a deposition. Her statements reinforced previous testimony from fellow ex-board member Helen Toner regarding trust issues with Altman and what they described as a toxic organizational culture.

 

McCauley characterized Altman’s leadership as creating chaos and crisis through what she called a culture of lying and deceit that extended throughout OpenAI’s leadership structure. She specifically cited an incident involving the launch of GPT4-Turbo, an AI model. According to McCauley, Altman falsely claimed that OpenAI’s legal department had determined the model did not require review by an internal safety board before its launch in India.

 

The former board member testified that Altman’s dishonesty resulted in crisis events occurring every few months. She referenced an email from Ilya Sutskever, another former OpenAI board member and influential executive within the company, which allegedly contained dozens of pages documenting chaotic events stemming from Altman’s behavior and alleged falsehoods.

 

David Schizer, a former Columbia Law School dean, appeared as an expert witness on nonprofit governance for Musk’s legal team. Despite the potentially dry subject matter, Schizer’s testimony proved significant to the case’s central questions about nonprofit law and organizational mission.

 

Working with Musk’s attorney Steven Molo, Schizer reviewed various actions taken by Altman that had been described by previous witnesses. Molo systematically asked whether these actions aligned with OpenAI’s safety-first mission and standard nonprofit practices. Schizer consistently indicated they did not.

 

One specific example involved complaints that OpenAI launched products without board knowledge, including allegations that Microsoft tested a version of GPT-4 without completing the company’s safety review process. Schizer emphasized the importance of board-CEO partnership in ensuring mission compliance. “The board and CEO need to be partnering, working together, to make sure the mission is being followed,” Schizer testified. “If the CEO is withholding that information, it’s a big problem.””