Elon Musk Faces Twitter Fallout During Court Testimony

On Wednesday, Elon Musk attended a federal court hearing in California, alleging that Sam Altman and his cofounders “stole a charity.” In a surprising twist, he admitted under oath that Tesla is not currently working toward artificial general intelligence, contradicting a tweet he made just weeks prior.

It was certainly a revealing day for Musk.

Musk’s lawsuit challenges the organizational structure of OpenAI, claiming that Altman and the others misled him into supporting a non-profit, which later transitioned to a for-profit model that now dominates the entity.

After several hours of testimony in the California courtroom, it appears the outcome may hinge on whether jurors and Judge Yvonne Gonzalez Rogers perceive a significant difference between capped and uncapped profits for OpenAI investors.

Musk explained that when he co-established the lab with Altman, Ilya Sutskever, Greg Brockman, and others, he initially trusted their vision for AI aimed at benefiting humanity. Over time, however, he grew wary of their intentions and eventually concluded that they were essentially “looting the nonprofit.”

Seeking to challenge Musk’s narrative during cross-examination, OpenAI’s attorney William Savitt pointed out Musk’s previous support for efforts to shift OpenAI toward a for-profit model to secure funding capable of competing with giants like Google, including the potential integration of the AI lab into Tesla itself.

Musk testified that he had contemplated converting OpenAI to a for-profit entity as early as 2016 and had even pursued plans in 2017 to maintain majority stake and control. He noted that after those arrangements fell through, he ceased regular donations to OpenAI but continued covering its office expenses until 2020.

Musk emphasized the crucial difference between investors with capped profits and those with unlimited returns. He noted that Microsoft’s initial investments in OpenAI featured profit limitations, but those constraints have since been relaxed. Musk attributes this evolution to his decision to sue.

Savitt sought to demonstrate that Musk was part of discussions led by Altman and Shivon Zillis—his longtime adviser—regarding fundraising efforts and did not raise any objections. Interestingly, Zillis also served on OpenAI’s board during the approval of those financial transactions.

The interrogation also touched on Tesla’s AI objectives. Musk faced scrutiny about his statements on Tesla’s AI ambitions. After declaring that the company’s focus was solely on self-driving technology, he was reminded of a recent tweet asserting, “Tesla will be one of the companies to make AGI.” In response, he stated under oath, “We are not pursuing AGI right now.” This revelation may cause concern for Tesla shareholders.

Musk was further queried about a tweet claiming he invested $100 million in OpenAI, contradicting the actual amount of $38 million. He contended that his reputation and connections compensated for this discrepancy.

Savitt referenced emails that showed Musk supported Tesla and Neuralink’s attempts to recruit staff from OpenAI while still serving on its board. Notably, one email discussed Andrej Karpathy’s transition from OpenAI to Tesla to lead self-driving initiatives, while another addressed Sutskever, whom Zillis suggested Musk hire.

A critical aspect of the day involved discussions around AI safety. Musk argued that OpenAI’s evolution into a traditional corporation poses a threat to public safety by diminishing its commitment to secure AI development. Savitt quickly pivoted Musk’s admission that all AI companies—including his own—face this risk.

Although Judge Gonzalez Rogers interrupted that line of questioning, she indicated that it would be revisited, albeit with some restrictions. When Musk’s legal team suggested probing ChatGPT’s involvement in the Tumbler Ridge shooting in 2024, she expressed reluctance to delve into controversies linked to AI but allowed for discussions about the safety measures of both xAI and OpenAI.

Musk is scheduled to appear again on Thursday for further questioning. Additional testimonies are anticipated, including those of his family office manager Jared Birchall, AI safety expert Stuart Russell, and OpenAI president Greg Brockman.