Inside OpenAI’s safety crisis: Former employees testify in Musk lawsuit

Elon Musk’s legal campaign to dismantle OpenAI’s for-profit structure is forcing a rare public examination of how the company’s shift toward commercial products may have compromised its founding mission: ensuring that artificial general intelligence (AGI) benefits all of humanity. On Thursday, a federal court in Oakland heard testimony from a former employee and a former board member who described a pattern of safety lapses and governance failures inside the AI lab.

Safety teams disbanded as product pressure mounted

Rosie Campbell joined OpenAI’s AGI readiness team in 2021 and left in 2024 after her team was disbanded. Another safety-focused group, the Super Alignment team, was shut down during the same period. Campbell testified that when she joined, the culture was heavily research-oriented, with frequent discussions about AGI and safety. “Over time it became more like a product-focused organization,” she said.

Under cross-examination, Campbell acknowledged that significant funding is necessary for building AGI, but argued that creating a super-intelligent model without adequate safety measures contradicts the mission she originally signed up for. She pointed to a specific incident where Microsoft deployed a version of OpenAI’s GPT-4 model in India through its Bing search engine before the company’s Deployment Safety Board (DSB) had evaluated it. While the model itself posed no major risk, Campbell stressed the importance of setting strong precedents. “We want to have good safety processes in place we know are being followed reliably,” she testified.

Board governance under scrutiny

The deployment of GPT-4 in India was one of the red flags that led OpenAI’s non-profit board to briefly fire CEO Sam Altman in November 2023. Tasha McCauley, a board member at the time, testified about concerns that Altman was not forthcoming enough for the board’s unusual structure to function effectively. She described a pattern of misleading behavior, including Altman lying to another board member about McCauley’s intention to remove a third board member, Helen Toner, who had published a white paper with implied criticism of OpenAI’s safety policies.

McCauley also noted that Altman failed to inform the board about the decision to launch ChatGPT publicly, and that his disclosure of potential conflicts of interest was inadequate. “We are a non-profit board and our mandate was to be able to oversee the for-profit underneath us,” she told the court. “Our primary way to do that was being called into question. We did not have a high degree of confidence at all to trust that the information being conveyed to us allowed us to make decisions in an informed way.”

When OpenAI’s staff rallied behind Altman and Microsoft worked to restore the status quo, the board reversed course, and the members opposed to Altman stepped down. This episode lies at the heart of Musk’s argument that the transformation of OpenAI from a research organization into one of the largest private companies in the world broke the implicit agreement among its founders.

Expert testimony and broader implications

David Schizer, a former dean of Columbia Law School who is serving as an expert witness for Musk’s team, echoed McCauley’s concerns. “OpenAI has emphasized that a key part of its mission is safety and they are going to prioritize safety over profits,” Schizer said. “Part of that is taking safety rules seriously, if something needs to be subject to safety review, it needs to happen. What matters is the process issue.”

With AI already deeply embedded in for-profit companies, the implications extend far beyond a single lab. McCauley argued that the governance failures at OpenAI should be a reason to embrace stronger government regulation of advanced AI. “If it all comes down to one CEO making those decisions, and we have the public good at stake, that’s very suboptimal,” she said.

Conclusion

The Oakland hearing underscores a fundamental tension at OpenAI: the pressure to commercialize AI products versus the non-profit mission of ensuring safe AGI. As Musk’s lawsuit proceeds, the testimony from former employees and board members is providing an unusually detailed look at how internal safety processes and governance structures have evolved—or failed to evolve—alongside the company’s rapid growth. For regulators, investors, and the public, the case is becoming a critical test of whether corporate accountability can keep pace with AI’s accelerating capabilities.

FAQs

Q1: What is the central issue in Elon Musk’s lawsuit against OpenAI?
The lawsuit argues that OpenAI’s shift from a non-profit research organization to a for-profit commercial entity violated its founding mission of developing AGI safely for the benefit of humanity. The court is examining whether this transformation broke implicit agreements among the founders.

Q2: What specific safety failures were highlighted in the testimony?
Former employee Rosie Campbell testified that the company’s Deployment Safety Board was bypassed when Microsoft deployed GPT-4 in India. She also noted that two key safety teams—the AGI readiness team and the Super Alignment team—were disbanded as the company became more product-focused.

Q3: How does this case affect the broader AI industry?
The case is being watched closely as a potential precedent for how AI companies balance safety and profit. Witnesses have called for stronger government regulation, arguing that relying on a single CEO to make decisions affecting public safety is “suboptimal.” The outcome could influence how other AI labs structure their governance and safety processes.