Sam Altman concedes he underestimated the backlash against OpenAI's Pentagon partnership, insisting governments must remain more powerful than AI companies.
Sam Altman has finally broken his silence on what he learned from OpenAI's controversial Pentagon deal, and his central takeaway is surprisingly candid: he misread the room. In a new interview with journalist Laurie Segall on the Mostly Human podcast, the OpenAI CEO admitted he "miscalibrated" the depth of public distrust toward both artificial intelligence and the federal government. It is a rare moment of self-reflection from a figure who has spent the past two years racing to deploy the most powerful AI systems in history while simultaneously warning about their risks.
The Pentagon deal, struck in February, saw OpenAI deploy its models on classified military networks. The partnership immediately drew intense criticism from civil liberties groups, technologists, and OpenAI's own user base. Protests persisted even after Altman pledged the technology would not be used for autonomous weapons or domestic surveillance. Speaking at his home with Segall, a journalist who has covered him for over a decade, Altman acknowledged that a vocal segment of the public simply does not trust the government to follow its own laws when given access to advanced AI tools.
"There's at least a group of loud people online who really don't trust the government to follow the law," Altman told Segall. "And that feels like a very bad sign for our democracy."
What makes Altman's admission significant is the question sitting underneath it. Who should control artificial intelligence? Altman dug his heels in on one answer: the state. He argued that one of the most consequential questions the world faces in the coming year is whether AI companies or governments will hold more power, and he firmly came down on the side of public institutions. "I think it's very important that the governments are more powerful," he said.
This position puts Altman at the center of an escalating debate in Silicon Valley. Companies like OpenAI, Anthropic, and Google DeepMind are currently making decisions that shape national security, economic infrastructure, and information ecosystems. As CNBC's analysis has made clear, the concentration of AI capabilities in a handful of private firms has alarmed regulators on both sides of the Atlantic. The European Union passed its AI Act in March, and the Biden administration's executive order on AI safety, issued in late 2023, was an early attempt to assert governmental oversight over the industry.
Altman pointed to historical precedents where government-led efforts produced transformative results: the Manhattan Project, the Apollo Program, the Interstate Highway System. His argument is that national security decisions should flow through democratically elected processes, not through the boardrooms of private labs. "Not me, and not the CEO of some other lab," he said.
But critics see a tension here. OpenAI's own corporate governance has been turbulent. The boardroom drama in November 2023, when Altman was briefly ousted and then reinstated, exposed how little external oversight existed over the company's decision-making. The Pentagon deal itself highlighted how quickly a private company can reshape military capabilities without meaningful public debate.
What This Means for Startups and the Market
For startups building AI products, Altman's comments signal a shifting landscape. The era of move-fast-and-break-things in AI is colliding with national security interests, and the companies that figure out how to work with government frameworks early will likely have an advantage. Defense contracts are becoming a significant revenue channel for AI firms. Palantir reported $2.2 billion in revenue in 2023, driven largely by government contracts, and its stock has surged as defense AI spending accelerates.
At the same time, public trust remains a fragile asset. OpenAI's brand took a measurable hit after the Pentagon deal. Developers and enterprise customers who built products on top of OpenAI's APIs raised questions about the company's direction and whether its safety commitments were genuine or performative. Altman's concession that he misread public sentiment suggests the company knows it has repair work to do.
Segall captured the core tension well when she told Business Insider that society is now grappling with a fundamental question: will artificial intelligence benefit everyone, or only a narrow few? The Pentagon deal forced that question into the open, and Altman's response reveals how even the most powerful figures in AI are still learning where the public draws the line.
Watch for two things in the months ahead. First, whether the U.S. Congress passes meaningful AI legislation that establishes clear boundaries for military and government use of AI systems. Second, whether other AI companies follow OpenAI's lead into defense or deliberately distance themselves from it to capture the segment of the market that remains deeply skeptical of military applications. The companies that navigate this tension most skillfully will shape the next phase of the AI industry.
