Mira Murati told the Court that Sam Altman was not always Honest

In a dramatic development in the ongoing federal court case in Oakland, California, former OpenAI executives Sam Altman and Mira Murati gave video depositions that provide new insights into the internal turmoil at the AI giant.

The case is now in its third week and is focusing on whether OpenAI’s original mission to create a non-profit has been ditched in favor of a pushing a hard sell on commercialization. Altman, OpenAI co-founder Greg Brockman, Satya Nadella and Microsoft played large roles in bringing the company towards a business-focused future, the suit alleges.

During her witness, Murati reportes the critical situation behind the scenes, At the close of year, there was a huge mess at OpenAI, The leaders were fighting, and the CEO, Sam Altman, was briefly fired and then brought back. One dramatic moment had her questioning if Altman would’ve been “candid, truthful, and honest” with her at the time. Altman paused for a moment and then said, “Not always.”

Murati also thought that Altman would “scapegoat” executives, pit them against one another and create rifts among the highest-ranking officials of the company, which she says would be a “toxic leadership culture. Her testimony has since made headlines and could have a huge impact on how one of the world’s most powerful AI companies moves forward.

The lawsuit also brings to light the controversial situation earlier this year when OpenAI’s board member fired Altman as CEO after rehiring him days later.It also sheds light on the controversial circumstance early this year, when board member of OpenAI removed Altman from his position as CEO and replaced him with another, only to call him back days later. This period of turmoil in leadership was brief and Murati served as interim CEO at the time.. She subsequently left OpenAI to found her own AI business called Thinking Machines Lab.

The lawsuit claims up to $150 billion in damages and wants to remove Altman and Brockman from OpenAI’s leadership, making it one of the biggest courtroom battles that the tech industry has witnessed in years..

A Lie About Safety and a Pattern of Deception

Murati’s most explosive deposition testimony involved a particular issue of the safety of AI. In a statement, Murati testified that Altman informed her that OpenAI’s legal team had decided that a new model of AI, described as “GPT-4 Turbo,” was not required to pass the company’s internal safety board for deployment before it was released. She had contacted OpenAI chief strategy officer Jason Kwon directly, and discovered that Altman’s account of what had happened was different from Kwon’s. Murati replied when asked if Altman was telling the truth, in the negative. She said she was concerned with a pattern, which would mean that Altman would give one person one answer and another person a completely different one. This made the work place very rough and turbulent, she said.

But Murati isn’t the only one working overtime to raise alarm bells. Former board members Helen Toner and Tasha McCauley testified that the former CEO’s honesty was just one data point among several that led to his firing. And before the decision was made to rehire Altman, co-founder Ilya Sutskever spent months gathering evidence of what he called “Altman’s pattern of lying,” a move he told me earlier this week he now regrets. On May 12, during an appearance before the board, Altman declared himself an honest and upright businessman who never meant to mislead the board.

TheTweaks Report:

TheTweaks Take on Murati’s testimony: It is more important than virtually anything else that was said in that courtroom. Not for the purpose of proving the legality of Musk’s case, but to establish what many in Silicon Valley have been saying all along. No one who says that, swears to tell the truth on oath, and that her very own CEO didn’t tell the truth, is talking gossip. This is a legally binding declaration. TheTweaks Take is that Altman will be able to live with the law, but the pride and trust in his closest collaborators can’t be recovered.

Now follow the instructions of the Trial with TheTweaks

Closing arguments will be Thursday May 14 and verdict could be as early as next week. The result may impact OpenAI’s future and could serve as a blueprint for the management of AI firms. TheTweaks is here to provide you with the latest news surrounding AI. Find the complete trial tracker here and stay updated about all OpenAI news and updates

For daily updates follow TheTweaks, the best news site that keeps you updated on AI. TheTweaks continues to update on the hottest issues and announcements of AI, tech and new releases.