He built the most consequential technology company of the decade. Whether the world can trust him to wield that power is genuinely disputed — by his own former colleagues, by the Economist, by the New Yorker, and by the structural logic of what he built.
In 2015, Sam Altman co-founded OpenAI as a nonprofit, framing it as a "Manhattan Project for A.I." — a safety-conscious counterweight to corporate AI development. By 2022, it had released ChatGPT, the fastest-adopted consumer product in history. By 2026, OpenAI had raised $122 billion at an $852 billion valuation.
In November 2023, OpenAI's nonprofit board fired him. Five days later, they reinstated him after nearly all of OpenAI's employees threatened to quit. The board members who fired him were replaced. The event — called internally "the Blip" — accelerated the dismantling of the governance structures meant to hold him accountable.
The question now is not whether Altman is powerful. He is. The Economist put it plainly: "In a very real sense, these five men hold the fate of Western civilisation in their hands." The question is whether the trust that power requires is warranted — and who gets to decide.
He promised a safety-first nonprofit. He built a for-profit empire at $852B. The New Yorker's April 2026 investigation found colleagues who believed he was "not trustworthy enough to have his finger on the button." The Economist: Altman "sought to portray himself as a bridge-builder" while doing the opposite in practice.New YorkerEconomist
The Economist concluded: "Should a handful of men be entrusted with the world's most potent new technology?" — and found America's laissez-faire approach "no longer politically tenable." The nonprofit board that could fire Altman was replaced with a compliant one. The for-profit conversion completed in 2025 removed the final structural brake.Economist
FT: "It's clear we won't regulate AI for safety's sake — as usual, governments will barely affect the trends that most affect our lives." Fortune found OpenAI has "consistently led the industry in subordinating safety concerns." Altman disputes this; argues no company has invested more in safety research.FTFortune
The most damaging account. Based on 100+ sources and internal documents. Found a pattern: saying one thing to regulators and the public, doing another internally. The board that fired him in 2023 had concluded he was "not consistently candid." Y Combinator's Paul Graham, who removed him from YC leadership, was "unambiguous that Altman was removed because of YC partners' mistrust."
The Economist's April 2026 leader — "America wakes up to AI's dangerous power" — concluded that the question of whether a handful of men should control the world's most potent technology demands a real answer. The status quo of minimal oversight is over. The Mythos moment (Anthropic's most capable model) showed what unregulated frontier AI deployment looks like.
Since the November 2023 coup, OpenAI has "consistently led the industry in subordinating safety concerns and rushing risky technologies to market — far more so than for-profit competitors like Microsoft and Google."
Internal accounts describe power concentrated around Altman personally following the failed 2023 coup, with safety research teams systematically overridden and internal objections suppressed.
The Economist's "Altman, Amodei and Musk fight dirty" piece found that "some admire him for getting the AI race going" while others see the bridge-builder rhetoric as cover for ruthless competition. Its key finding: the three AI CEOs have created a commercial race that may force anyone to compromise safety, regardless of intention. OpenAI's $13B revenue in 2025 is the largest — and the pressure it creates is structural, not personal.
The Economist's quantitative study of tech tycoon power: Altman currently falls in the bottom half of historic technology tycoons by measurable power — but "the technology they wield is still only in its infancy." The question of whether he becomes a Rockefeller figure is open, not settled. This is the most important reason the trust question is not yet resolved: the stakes are still arriving.
Benjamin Wallace-Wells's review of two major books concludes that the AI race was not inevitable — but that framing Altman as a supervillain "risks inflating contingent figures out of ambient anxiety." Anthropic, founded by OpenAI's safety-focused defectors, "would in time show little divergence" from OpenAI's model. The system rewards scale above all else; individuals operate within it.
The FT reported that Altman is refocusing OpenAI as Anthropic tests its early enterprise lead. The valuation is under pressure not because of trust concerns but because the commercial thesis — OpenAI as the dominant enterprise AI platform — is being contested. The trust question and the commercial question are converging.
Hao: "Everything OpenAI did was the opposite of inevitable." But Anthropic, founded by safety-focused OpenAI defectors, built the same thing. The original sin was structural — the term "artificial intelligence" itself, coined in 1955, set up decades of hype that made Altman's model possible. He surfed a wave; he didn't create it.
Warren's formal letter to OpenAI demanded answers on the for-profit conversion and whether the nonprofit's original mission had been betrayed. The first serious Congressional pressure on OpenAI's governance — a signal that the political environment around AI accountability is shifting.
Wallace-Wells argues Nadella's influence over OpenAI is "at least as consequential as Altman's." Nadella intervened to reinstate Altman in 2023 and seems "broadly allergic to boomer-doomer metaphysics" — his instinct is to narrow every philosophical question into project architecture. This may be reassuring. Or it may mean the accountability question is being deflected rather than answered.
"Some admire him for getting the AI race going. In contrast to Musk, Altman sought to portray himself as a bridge-builder — but his actions told a different story."
The New Yorker (twice), Fortune, and Ars Technica have published major sceptical investigations. The Economist (three articles) presents a more structural analysis: the problem is partly Altman, partly the system. The FT tracks the commercial consequences. Together they represent the most concentrated critical scrutiny any tech CEO has received since Mark Zuckerberg in 2018.
The EU AI Act treats AI governance as a regulatory matter, not a question of individual character. European regulators are less interested in whether Altman personally is trustworthy and more interested in whether the institutions governing OpenAI are adequate. Their answer: no — and the Mythos moment (Anthropic's most capable uncontrolled model) reinforced this.
Karen Hao quotes a Chinese AI researcher: "In China, no team of researchers would get $1 billion without an articulated vision of exactly what it would be good for." The framing in Asia and the Global South is less about Altman's character and more about whether US AI governance is adequate for a technology that affects everyone — and whether anyone is asking the countries most affected.
The trust question is real. The Economist put it plainly: "In a very real sense, these five men hold the fate of Western civilisation in their hands." Whether Altman specifically is trustworthy to hold that position is genuinely contested — by the New Yorker's two major investigations, by the Economist's structural analysis, by Fortune's internal reporting, and by the former colleagues who tried to remove him.
But Wallace-Wells in the New Yorker adds the most useful corrective: framing Altman as a supervillain "risks inflating a contingent figure out of ambient anxiety." Anthropic, founded by safety-focused OpenAI defectors, built essentially the same thing. The structural logic of the race may be stronger than any individual's intentions.
The honest summary: Altman has been less transparent than he promised. The governance structures designed to hold him accountable were dismantled after the 2023 coup. Safety concerns have been subordinated to commercial imperatives, per multiple credible investigations. The Economist is right that America's laissez-faire approach is no longer tenable. The FT is right that his commercial position is being tested by Anthropic.
What most needs to change is not Altman personally — it is the accountability architecture around him. That is where the Economist, the New Yorker, and Senator Warren all converge: the question of whether he can be trusted is less important than ensuring that trust is not required in the first place.