Murati Turns on Altman in Court: OpenAI Power War Exposed

Former CTO Mira Murati Breaks Silence, Accuses Altman of Deception

Mira Murati, the former Chief Technology Officer of OpenAI, has delivered explosive courtroom testimony directly accusing CEO Sam Altman of deliberately misleading her on model safety approvals and systematically creating 'chaos and distrust' among the company's senior leadership. The testimony, delivered via video deposition during Elon Musk's ongoing lawsuit against OpenAI, marks the most damaging insider account yet of how the world's most influential AI company operates behind closed doors.

This is not an outsider's attack. Murati was the technical backbone of OpenAI during its most consequential period — the launch of ChatGPT and the development of GPT-4. Her sudden departure in 2024 sparked widespread speculation, but she remained silent. Until now.

Key Takeaways From Murati's Testimony

• Safety process manipulation: Murati alleges Altman told her that OpenAI's legal team said the company's Safety Advisory Group did not need to review a new model before release — a claim she later discovered was not accurate.
• Information compartmentalization: Different executives allegedly received different versions of the same facts, preventing anyone from assembling a complete picture of critical decisions.
• Deliberate confusion: Murati characterizes Altman's management style as one designed to concentrate final decision-making power in his own hands by keeping others off balance.
• Broken trust: The testimony paints a portrait of an organization where senior leaders could not rely on the information they were being given by their own CEO.
• Timing significance: The revelations come as OpenAI pursues a controversial restructuring from its nonprofit origins to a for-profit entity, a move Musk's lawsuit seeks to block.

Inside Altman's Alleged 'Information Architecture'

Murati's most striking allegation concerns the way safety approvals were handled for a new model release. According to her testimony, Altman told her that Jason Kwon, who led OpenAI's legal affairs, had determined that the company's internal Safety Advisory Group did not need to sign off on the model before it went public.

Murati says she took Altman at his word. But the implication of her testimony is clear: this information was either inaccurate or deliberately misleading. The result was that a critical safety checkpoint was bypassed — not through a formal policy change, but through what amounted to a verbal shortcut from the CEO.

This pattern, Murati suggests, was not an isolated incident. She described a management approach built on selective information disclosure. Key facts were parceled out to different executives in different configurations. No single leader, aside from Altman himself, ever had the full picture.

The effect was a kind of engineered dependency. Every major decision ultimately funneled back to one person, because only one person held all the pieces.

Why Murati's Words Carry Unique Weight

Unlike many critics of OpenAI who speak from the outside, Murati's credibility stems from her central role in the company's most important achievements. She joined OpenAI in 2018 and rose to become CTO, overseeing the technical development that turned the research lab into a global phenomenon.

• She was the public face of the GPT-4 launch in March 2023
• She led technical operations during ChatGPT's explosive growth to over 100 million users
• She briefly served as interim CEO during Altman's dramatic 5-day ouster in November 2023
• She departed OpenAI in September 2024 with little public explanation

Her silence after leaving was conspicuous. Many in the industry assumed she had signed extensive non-disclosure agreements or simply wanted to move on. The courtroom, however, offered a venue where silence was no longer an option — and where legal protections for testimony applied.

The fact that someone of Murati's stature is now on the record accusing Altman of systematic deception fundamentally changes the narrative around OpenAI's internal governance.

The Musk Lawsuit Provides the Stage

Elon Musk originally co-founded OpenAI in 2015 as a nonprofit dedicated to developing artificial general intelligence (AGI) safely and for the benefit of humanity. He departed the board in 2018 but filed suit in 2024, alleging that Altman and co-founder Greg Brockman had betrayed the organization's founding mission by pivoting toward a profit-driven model closely tied to Microsoft, which has invested over $13 billion in the company.

Musk's legal team has argued that Altman engineered OpenAI's transformation through a series of calculated moves designed to consolidate his personal control while enriching himself and select insiders. Murati's testimony slots directly into this narrative, providing an insider's account of how information control allegedly functioned as a governance mechanism.

The trial, taking place in federal court, has already produced other notable moments — including internal communications and emails that paint a complex picture of OpenAI's evolution. But Murati's video deposition stands out for its directness and the authority of the witness.

Compared to previous whistleblower accounts from more junior employees, Murati's testimony hits at the C-suite level, making it far harder for OpenAI to dismiss.

OpenAI's Governance Crisis Deepens

Murati's testimony arrives at an extraordinarily fragile moment for OpenAI's organizational structure. The company is in the middle of a complex and contested restructuring plan that would convert its unusual nonprofit-controlled structure into a more conventional for-profit corporation.

This restructuring has drawn scrutiny from multiple directions:

• State attorneys general have raised concerns about whether the nonprofit's assets are being properly valued and protected
• Musk's lawsuit directly challenges the legality and ethics of the conversion
• Former board members have expressed reservations about the speed and terms of the transition
• Safety-focused researchers who have departed the company — including Ilya Sutskever, Jan Leike, and others — have raised alarms about the deprioritization of safety work
• Microsoft itself has had to navigate a delicate relationship as both OpenAI's largest investor and a potential competitor

The picture that emerges is one of an organization whose governance mechanisms have been under severe strain for years — and where the CEO's management style may have been a contributing factor rather than a stabilizing force.

What This Means for the AI Industry

The implications of Murati's testimony extend well beyond OpenAI's internal politics. The company is arguably the single most important organization in the current AI landscape. Its models power products used by hundreds of millions of people. Its technical decisions set the pace for the entire industry. And its approach to safety — or lack thereof — influences how competitors like Google DeepMind, Anthropic, Meta AI, and others calibrate their own risk frameworks.

If the world's leading AI company was bypassing its own safety review processes through informal executive communications, that raises profound questions about the adequacy of self-governance in the AI sector.

For developers and businesses building on OpenAI's APIs, the testimony introduces a new dimension of risk. The reliability of an AI provider depends not just on technical uptime but on organizational stability and trustworthy governance. Companies making long-term bets on OpenAI's ecosystem may now factor governance risk more explicitly into their planning.

For policymakers in Washington and Brussels, Murati's account provides ammunition for those arguing that voluntary safety commitments from AI companies are insufficient. The EU AI Act is already moving toward mandatory compliance frameworks, and testimony like this could accelerate similar efforts in the United States.

Looking Ahead: The Trial's Ripple Effects

The Musk v. OpenAI trial is expected to continue producing revelations in the coming weeks. Murati's testimony sets a precedent — other former executives may feel emboldened or compelled to share their own accounts.

Several critical questions remain unresolved:

What specific model was involved in the safety review bypass Murati described? How did Jason Kwon respond to the characterization of his alleged legal guidance? Will Altman himself take the stand and address these allegations directly? And will the court's eventual ruling have any binding effect on OpenAI's restructuring plans?

What is already clear is that the mythology of OpenAI — the story of a mission-driven organization racing to build safe AGI for humanity — has suffered a blow from which it may not recover. When your own former CTO testifies that the CEO built a management system on 'chaos and distrust,' the narrative of benevolent innovation becomes very hard to sustain.

Sam Altman has survived boardroom coups, public controversies, and relentless competition. But Murati's testimony represents a different kind of threat: a credible insider, with nothing obvious to gain, calmly describing a pattern of deception under oath. This may be Altman's most dangerous battle yet — and the outcome will shape not just OpenAI's future, but the governance standards for the entire AI industry.