Murati Testifies She Could Not Trust Altman

Mira Murati, OpenAI's former Chief Technology Officer, has testified under oath that CEO Sam Altman lied to her about the safety clearance for a new AI model. The bombshell video deposition, shown during the ongoing Musk v. Altman trial on Wednesday, adds explosive new detail to one of the most consequential legal battles in the history of artificial intelligence.

Murati stated that Altman falsely told her OpenAI's legal department had determined a new AI model did not require additional safety review before deployment. Her testimony paints a troubling picture of internal governance at the world's most prominent AI company — and raises fresh questions about whether commercial pressures are overriding safety protocols at OpenAI.

Key Takeaways From Murati's Testimony

• Altman allegedly misrepresented legal clearance: Murati said Altman told her OpenAI's legal team had approved a new model for deployment without further safety review — a claim she says was false.
• Trust breakdown at the top: The former CTO explicitly stated she reached a point where she 'couldn't trust' Altman's words on critical safety matters.
• Testimony occurred under oath: Murati's statements carry legal weight as a sworn video deposition in the Musk v. Altman federal trial.
• Timing aligns with leadership turmoil: Murati's departure from OpenAI in September 2024 came after the chaotic board upheaval of November 2023.
• Broader safety concerns surface: The testimony reinforces a pattern of allegations that OpenAI has deprioritized safety in its race to commercialize AI products.
• Elon Musk's legal team leverages the testimony: The deposition is being used to support Musk's claims that OpenAI has abandoned its original nonprofit, safety-first mission.

Inside the Courtroom: What Murati Actually Said

The video deposition was played for the jury during the third week of the Musk v. Altman trial in federal court. Murati, who served as OpenAI's CTO from 2022 until her resignation in September 2024, described a specific incident in which Altman communicated information about a model's safety clearance that she later discovered to be inaccurate.

According to Murati's testimony, Altman told her that OpenAI's internal legal department had reviewed a new AI model and concluded it did not need to go through additional safety evaluation before being released. Murati said she later learned this was not the case — the legal team had not made such a determination.

The revelation is significant because it suggests a disconnect between what OpenAI's leadership communicated internally and what its safety and legal teams actually concluded. For a company that has repeatedly promised the public it takes AI safety 'extremely seriously,' the allegation strikes at the heart of its credibility.

Why This Testimony Matters for the Musk v. Altman Trial

Elon Musk filed his lawsuit against Sam Altman and OpenAI in early 2024, alleging that the company had betrayed its founding mission as a nonprofit dedicated to developing artificial general intelligence (AGI) safely and for the benefit of humanity. Musk, who co-founded OpenAI in 2015 and contributed over $44 million to the organization, argues that Altman orchestrated a transformation into a profit-driven entity — culminating in its partnership with Microsoft, which has invested approximately $13 billion in the company.

Murati's testimony directly supports one of Musk's central arguments: that OpenAI's leadership has systematically sidelined safety considerations in favor of speed and commercial advantage. If the jury finds her account credible, it could significantly bolster Musk's case.

The trial has already featured testimony from multiple former OpenAI employees and board members. But Murati's deposition carries particular weight because of her seniority — as CTO, she was one of the most technically authoritative voices inside the company and had direct oversight of model development and deployment decisions.

A Pattern of Leadership Turmoil at OpenAI

Murati's testimony does not exist in a vacuum. It fits into a well-documented pattern of internal conflict and leadership instability at OpenAI over the past 2 years.

• November 2023 board crisis: OpenAI's board of directors fired Sam Altman as CEO, citing concerns that he was not 'consistently candid' with the board. Altman was reinstated just 5 days later after a massive employee revolt and pressure from Microsoft.
• Ilya Sutskever's departure: Co-founder and former Chief Scientist Ilya Sutskever, who initially supported the board's decision to fire Altman, left OpenAI in May 2024 to found a new AI safety company called Safe Superintelligence Inc.
• Murati's resignation: Murati herself departed in September 2024, just months after being named interim CEO during the November crisis. Her exit was one of several high-profile departures that stripped OpenAI of senior technical leadership.
• Jan Leike's exit: Jan Leike, co-lead of OpenAI's Superalignment team focused on long-term AI safety, resigned in May 2024 and publicly stated that 'safety culture and processes have taken a backseat to shiny products' at the company.

Taken together, these departures represent a significant brain drain of safety-focused leadership from OpenAI. Murati's courtroom testimony now provides sworn, legally binding confirmation of the internal tensions that drove these exits.

The Broader AI Safety Debate Intensifies

Murati's allegations arrive at a critical moment for the global AI safety conversation. Governments around the world are grappling with how to regulate increasingly powerful AI systems, and OpenAI has positioned itself as a leader in responsible AI development.

The company published its Preparedness Framework in late 2023, outlining a structured process for evaluating the risks of new models before deployment. OpenAI has also been a vocal participant in international AI safety summits, including the AI Safety Summit at Bletchley Park in the United Kingdom and subsequent gatherings in Seoul and Paris.

But if Murati's testimony is accurate, it suggests that OpenAI's public commitments to safety may not always translate into internal practice. This gap between rhetoric and reality is precisely what critics — including Musk, former employees, and AI safety researchers — have warned about.

Compared to competitors like Anthropic, which was founded by former OpenAI researchers specifically to prioritize safety, OpenAI now faces heightened scrutiny over whether its governance structures are robust enough to prevent commercial incentives from overriding safety protocols. Anthropic's Responsible Scaling Policy and its Constitutional AI approach have been widely praised as more transparent and rigorous than OpenAI's framework.

What This Means for Developers, Businesses, and Users

The implications of Murati's testimony extend far beyond the courtroom. For the millions of developers and businesses building on OpenAI's APIs and products, trust in the company's safety claims is foundational.

• Enterprise customers who rely on OpenAI's models for sensitive applications — healthcare, finance, legal — need assurance that models undergo rigorous safety testing before release.
• Developers integrating GPT-4, GPT-4o, and future models into their products may face increased pressure from regulators and users to verify safety claims independently.
• Investors in OpenAI, which was valued at approximately $300 billion in its most recent funding round, must weigh reputational risk alongside growth projections.
• Policymakers in the U.S. and EU may use the testimony as evidence that voluntary self-regulation by AI companies is insufficient, potentially accelerating legislative action.

For everyday users of ChatGPT — which now serves over 200 million weekly active users — the testimony raises uncomfortable questions about whether the products they rely on daily have been subjected to the safety rigor OpenAI publicly promises.

Looking Ahead: The Trial's Next Steps and Industry Impact

The Musk v. Altman trial is expected to continue for several more weeks, with additional witnesses and depositions likely to surface more internal details about OpenAI's decision-making processes. The outcome could have far-reaching consequences.

If the court rules in Musk's favor, it could potentially force OpenAI to restructure its corporate governance, limit its for-profit activities, or even unwind aspects of its Microsoft partnership. A ruling against Musk, on the other hand, would largely validate OpenAI's current trajectory.

Regardless of the legal outcome, Murati's testimony has already achieved something significant: it has placed the question of AI safety governance squarely in the public record, backed by sworn statements from one of the industry's most credible technical leaders. The AI industry will be watching closely as the trial unfolds — and so will regulators in Washington, Brussels, and beyond.

OpenAI has not yet issued a detailed public response to Murati's specific claims. The company has previously maintained that it follows rigorous internal safety processes and that its models undergo extensive evaluation before deployment. Sam Altman has not commented publicly on the deposition.

One thing is clear: the era of AI companies self-certifying their own safety standards may be drawing to a close. And Mira Murati's 4 words — 'I couldn't trust him' — may prove to be the catalyst.