A federal courtroom in Oakland turned into a window into OpenAI’s turbulent inner workings this week, as former chief technology officer Mira Murati described a leadership culture she said was marked by confusion, mistrust and competing narratives under CEO Sam Altman.
Murati, appearing through recorded testimony in Elon Musk’s lawsuit against OpenAI, told the court that Altman often delivered sharply different messages to different executives, creating friction at the highest levels of the company.
According to her testimony, the inconsistency was not occasional — it became a pattern that unsettled senior leadership during one of the most consequential periods in OpenAI’s rise.
She said the environment frequently felt chaotic, adding that she and others struggled to understand where decisions truly stood from one conversation to the next.
Murati briefly stepped into the CEO role during the dramatic 2023 boardroom upheaval that temporarily removed Altman from power. Despite her concerns, she told the court she still believed keeping him at the helm was important, largely because the company itself appeared dangerously unstable during the crisis.
At one point, she warned, OpenAI seemed close to fracturing entirely.
The testimony forms part of Musk’s sprawling legal challenge accusing OpenAI of abandoning the nonprofit principles on which it was founded in favor of aggressive commercial expansion. Musk, who helped launch OpenAI before later breaking away, is seeking massive financial damages reportedly valued at $150 billion, with claims tied to the organization’s structural shift and its partnership with Microsoft.
The case has evolved into more than a dispute over corporate governance. It has become a public dissection of how one of the world’s most influential AI companies transformed from a research-focused nonprofit into a dominant commercial force shaping governments, classrooms and global business.
Another former OpenAI insider, Shivon Zilis, also testified about internal anxiety surrounding the release of ChatGPT, the chatbot that catapulted the company into mainstream prominence.
Zilis said members of the board expressed serious concern over how quickly the product was being pushed out and whether directors were adequately informed before launch decisions were made.
Without detailing every disagreement, she acknowledged there had been prior concerns raised internally about Altman’s conduct and leadership style.
The trial has steadily revealed long-simmering tensions between founders and executives over power, governance and the future direction of artificial intelligence development. Testimony so far has touched on Musk’s early financial backing of the startup, disputes over who should lead the company and even a failed attempt to resolve conflicts privately before proceedings began.
What is emerging from the courtroom is less a story about code and algorithms — and more one about ambition, loyalty and control inside the company now sitting at the center of the global AI race.
