How ‘Effective Altruism’ went from fringe philosophy to Trump’s AI villain

“Americanism, not effective altruism,” the Department of War tweeted in mid-Sept. “The United States will continue to be AI DOMINANT!” Defense Secretary Pete Hegseth chastised Effective Altruism for “sanctimonious rhetoric” used to “strong arm the United States military into submission.” Emil Michael, the Pentagon’s Under Secretary of Defense for Research and Engineering, called its ideology “dangerous” and “pretty dystopian.”

But what is “Effective Altruism” and how did this once obscure philosophical movement wind up in the crosshairs of some of the most powerful U.S. officials and their efforts to argue against AI regulation? 

Founded by two British academics almost two decades ago, Effective Altruism is a philosophical movement whose core idea is that people should try to use rational processes to do the most good for the most people in the world. That may not sound controversial, but as with so many dogmas that purport to have simple, good intentions, things get messy when those ideas encounter the complexity of the real world—and competing ideas about what people and governments should prioritize and why.

Effective Altruism, or EA for short, has gotten caught up in the AI debate because many of those who subscribe to its philosophy have gravitated towards working at the top AI companies: OpenAI, Anthropic, and Google DeepMind. They have done so because of a fervent belief in both AI’s possibility to advance humanity, and its potential to cause great harm. In particular, many EA adherents are drawn to work on AI safety, believing that unless methods are found to ensure advanced AI systems remain obedient to human instructions and human values, this powerful technology has a good chance of destroying humanity.

EA’s close affiliation with fears about advanced AI leading to human extinction—often branded “AI doomerism”— has brought the group into conflict with the Trump administration, which is racing to keep AI regulation off the table over fears of ceding ground to China in the AI race and concerns that any slowdown in the pace of AI development could tank the U.S. stock market, which has been buoyed by the runup in shares of AI-related stocks.

Despite loud warnings from industry leaders including OpenAI’s Sam Altman and Anthropic’s Dario Amodei last week, Trump was quick to reject the idea of an international slowdown. For Trump and other opponents of regulation, using culture war rhetoric to attack EA has become a way to try to discredit the broader anxiety about AI risk before it hardens into legislation. As public concern about such risks reach a fever pitch, some worry the attacks on EA could rupture the political common ground that’s quietly starting to form on AI rules.

The only problem? Since the administration kicked off its anti-AI doomerism social media campaign, taking particular aim at EA and its adherents’ belief that AI carries a high risk of ending humanity—anti-AI sentiment seems to have grown. A recent poll from Politico found that 63% of Americans believe that AI poses at least a moderate risk of destroying humanity.

Which raises the question: Will branding EA an enemy of America neutralize it—or give it the biggest platform it’s ever had?

What is effective altruism?

EA is in some ways a textbook political target for Trump. Its members are mostly left-leaning (70%), according to an 2024 survey conducted by the Effective Altruism Forum. Some of its members have advocated for open borders, and believe gender is a social construct, both of which have been fodder for inflammatory social media content from those seeking to discredit the philosophy.

As for other demographics, the group is mostly male (68.8%), white (75%), young (average age 32.4 years), and highly educated—a fact that makes it easy to caricature as an elitist and esoteric movement out of touch with reality.

EA traces back to 2009, when Oxford philosophers Toby Ord and William MacAskill founded Giving What We Can, a pledge for members to donate at least 10% of their income to the most effective charities. The idea was to treat charity and ethics like a math problem, figuring out which causes saved the most lives, or did the most good, per dollar spent. 

Lawrence Sumulong—Getty Images

Those drawn to the philosophy tend to be interested in “quantitative forms of reasoning,” said Conor McGlynn, a PhD candidate at Harvard who works on AI policy and the intellectual history of AI safety.

A core idea of EA is that all human lives were of potentially equal value. Given this, the founders of EA argued, it is often best to spend money in the developing world, where simple interventions—such as buying bed nets to ensure children did not contract malaria—can save far more lives than donating an equivalent sum on something like a homeless charity in the U.S. or Europe.

Another core conclusion of its analytics-based approach to charity is that capital is philanthropically more powerful than labor. By directly volunteering at a local soup kitchen three hours a week, a person might only be able to help a few dozen needy people at a time. But if that person were, say, a lawyer who could earn $500 per hour, then donating three hours’ worth of wages to a charity that bought anti-malarial bed nets could save several hundred people’s lives every week. 

By this logic, it might be better for most people to try to earn as much money as possible and then donate most of it to charity, rather than trying to do good works directly. EA called this idea “earn to give.”

In 2011, to further this, MacAskill launched 80,000 Hours, a career-advice project steering talented people toward high-impact work. A year later, it folded the project and “Giving What We Can” under a new umbrella group, the Centre for Effective Altruism (CEA). It operated under a parent organization called Effective Ventures until spinning out as its own independent nonprofit in 2025.

CEA has published and periodically revised a set of “guiding principles” defining the movement’s core commitments—among them a “Commitment to Others” (“willing to take significant personal action” to benefit others), a “Scientific Mindset” (basing actions on evidence and reasoning rather than intuition), and “Openness” (staying tied to a method rather than a fixed cause, and changing course if the evidence points elsewhere). The group also hosts events such as the annual EA Global conference series, maintains an active online forum, and conducts a periodic survey, but it is not strictly a membership organization. 

EA’s overlap with AI safety

EA’s involvement with AI risk grew out of the movement’s parallel interest in “longtermism”—the idea that positively shaping the far future is a moral priority. 

This interest came from an EA belief that the lives of future people are as valuable as those alive today. Since far more people have yet to be born than currently exist, the longtermist argument is that preventing human extinction would do more good than almost any other cause. The link with AI found new force after philosopher Nick Bostrom’s 2014 book Superintelligence, which argued AI posed a singular existential threat, and Toby Ord’s 2020 book The Precipice, which singled out unaligned AI as an existential risk.

“There was a thought that AI was coming, and it was going to be wildly influential, hugely transformative, and that actually it was going to be incredibly important from the perspective of doing the most good that we can to steer the development of AI in the direction towards good and away from harm,” said according to Brian Hedden, a professor of linguistics and philosophy at the Massachusetts Institute of Technology (MIT). 

Because of this moral imperative, many EA-aligned individuals and charitable organizations have worked to address these risks, with many offering financial aid to further AI safety research. Open Philanthropy, an EA-aligned charity which counted tech billionaire Dustin Moskovitz among its cofounders, steered significant money toward AI safety research over the last decade, a shift that accelerated as large language models advanced. Another cofounder was Holden Karnofsky, who is now an Anthropic employee and married to Anthropic cofounder and president Daniela Amodei. 

Another prominent EA figure and AI safety research bankroller was Sam Bankman-Fried. The disgraced crypto founder was one of the movement’s most prominent adherents and its biggest donors, building crypto exchange platform FTX into a multibillion-dollar company with the explicit goal of amassing a fortune to give away.

After it emerged that Bankman-Fried had misappropriated billions in FTX customer funds and Bankman-Fried was convicted of fraud, his downfall became a parable of the movement’s own utilitarian logic taken to its extreme—proof that “earn to give” could curdle into a license to cut ethical and legal corners.

While the EA movement has historically been one of the only groups taking seriously the prospect of AGI—a type of artificial intelligence that surpasses all humans—and the extreme risks it could pose, not all effective altruists are AI doomers. In fact, there are many “different flavors of EA,” Hedden said. “It’s painting with an overly broad brush to say that effective altruism is opposed to further AI development. 

EA-aligned groups fund safety causes  

EA’s deep ties to AI labs, safety research, and regulatory efforts have led some in the Trump administration to accuse the movement of having too much influence over the wider industry.

Anthropic has emerged as a particular target, since many of its cofounders either have ties to EA or have espoused views aligned with its philosophy—and the company received some of its initial funding from EA-linked investors. (Bankman-Fried was an early backer of Anthropic, leading the company’s Series B funding round.) The lab has also attracted more EA aligned employees due to its founding focus on mitigating AI risk. That said, Anthropic does not publicly align itself with the EA movement.

In February, Pete Hegseth referenced effective altruism in a tweet attacking Anthropic’s stance on not allowing the U.S to use its technology for military operations. “Cloaked in the sanctimonious rhetoric of ‘effective altruism,’ they have attempted to strong-arm the United States military into submission—a cowardly act of corporate virtue-signaling that places Silicon Valley ideology above American lives,” he wrote.

The Pentagon’s Emil Michael, and David Sacks, the former AI and crypto czar, have also been vocal critics of EA’s influence over the current AI safety landscape. Both have also lobbied against AI regulatory efforts. Last week, in an interview with Fox News, Michael said of EA: “They put themselves on par with the president and other democratic presidents,” he said.

Tierney L. Cross—Bloomberg via Getty Images

“For some people within the White House, the AI safety community and the effective altruists have always been an enemy,” Andrew Kane, policy director at the Secure AI project, said. Kane was previously chief of staff at the National Institute of Standards and Technology (NIST), an agency of the United States Department of Commerce. 

“While maybe [AI safety] was a niche issue three years ago that was more EA focused—it’s becoming more apparent to the public, and it’s expanded beyond that,” he added. “They are trying to pin this on some shadow group of people to distract from the fact that this is becoming a very real public concern that is much broader than that.”

Part of the reason the group has found itself under attack is that it has been involved with pro-regulatory policy groups. The EA-backed Open Philanthropy Action Fund spent $110,000 in the second quarter of 2026 on lobbying efforts. Another organization, the Center for AI Policy, is an advocacy organization within the EA movement that pushed for U.S. federal legislation to mitigate catastrophic risks from AI has been active since 2024, when it registered its first lobbyist and promptly spent $98,000 in one month, Politico reported in 2024. 

Many in the AI safety world argue that while these funding and political ties to EA are real, it doesn’t mean everyone concerned with AI risk is subscribed to the ideology. Some researchers say the group has found itself the target of such attacks because it’s a small, identifiable network with outsized financial and personnel ties to the very labs building the technology—making it easier to cast as a shadowy special interest than to engage with the broader public unease about AI.

Seán Ó hÉigeartaigh, a research professor at the Centre for the Future of Intelligence,who has taken money from EA-aligned funders himself, said the EA community is very involved with AI safety research by virtue of the fact it took the issue seriously before many. While it’s hard to rule out subtler effects, such as researchers shaping a proposal to fit what a funder is likely to back, he said, there hasn’t been any attempt to implement message control.

Early involvement also isn’t the same as ownership, Tyler Johnston, founder of AI watchdog group the Midas Project who identifies with the effective altruist movement, said. He compared the current debate to questioning if anyone concerned about climate change is involved in the environmentalist movement. EA’s early concern about AI safety, he said, produced both “serendipitous” and deeper connections among the people now working in safety. 

Trump’s AI agenda: down with the doomers

For now, Trump’s priorities are firmly focused on his AI accelerationist and anti-AI slowdown agenda. In a post last week, he shared a post claiming the only safeguard needed to control AI was a “high IQ president.”

One side effect of this is a new distance from OpenAI, in addition to its existing chilly relationship with Anthropic, and from the growing number of AI employees in Silicon Valley who also support the calls for a slowdown. It’s a notable change from January 2025, when Altman was currying favor with Trump, saying he would do “incredible” things for the U.S. and standing alongside him to announce the administration’s $500 billion Stargate investment in AI infrastructure. 

Trump has instead been forming tighter alliances with those aligned with this anti-doomer worldview, mostly notably with Nvidia CEO Jensen Huang. Last week, Huang answered a call from Trump while he was on stage, and put the President on speaker phone. Through the microphone, Trump said to the crowd, “Whatever Dario said this weekend won’t stop our progress.” He also called the backlash to data centers and fears that AI will “take over” a “hoax.”

Still, the public pushback coming from figures like Sacks and Michael diverges from many in the Republican Party, according to Kane. 

“In Congress, Republicans who two years ago were just trying to ban all state regulation are now saying, actually, we need to draft something. That was unheard of a year ago,” he said. The mood, both among the public and within the government, seems to be tipping toward taking AI risk seriously and wanting at least some regulation, he said.

“It’s odd that we’re seeing this pushback focused on EA” he said. “But it’s very out of step with what the conversation actually is from a lot of policymakers, on both sides.”