
MELBOURNE, Australia — “Americanism, not effective altruism.” That’s a new slogan launched on X by the United States Department of “War,” the words running across an image of President Donald Trump saluting. Above the image, the post reads: “The United States will continue to be AI DOMINANT!” I initially thought it must be satire. Doesn’t the defense department have more important things to do than oppose a movement that encourages people to help others as effectively as possible? How could an administration that created a Department of Government Efficiency with great fanfare now oppose being effective? It’s less than 20 years since I spoke with the small group of students at the University of Oxford who started effective altruism (EA). Influenced by my 1972 article “Famine, Affluence, and Morality,” they were working together to put into practice what I had advocated, adding an emphasis on getting the best value for every pound or dollar they could donate. They discovered that some charities are hundreds of times more effective than others, and posted their research online so that everyone could use it.
Giving What We Can, the first EA organization, asked its supporters to pledge to donate at least 10 percent of their income to highly effective charities. GiveWell says donors have directed more than $2.6 billion to their recommended organizations, and estimates that these donations will save more than 340,000 lives. I founded The Life You Can Save, which, to date, has raised more than $140 million for highly effective charities working in healthcare, education, and overcoming poverty.
Since its founding, EA has taken on issues other than global poverty. The suffering we inflict on animals matters, too, and many EAs seek to reduce that suffering, often drawing on research from animal charity evaluators to find the most effective organizations to support.
Other EAs became concerned about preventing catastrophic risks, especially those, like nuclear war or bioterrorism, that could cause the extinction of our species. They took seriously the possibility that superintelligent artificial intelligence (AI) could eliminate its human creators — something that Alan Turing, the father of computer science, warned about in the 1950s. Some EAs argued that the loss of all future human generations would be a catastrophe far outweighing all short-term concerns, and that preventing it should be the highest priority.
It is this aspect of EA that aroused the Trump administration’s opposition. We can start with OpenAI’s disclosure in July that AI agents it was testing had escaped their sandbox, accessed the internet, and hacked into another company’s website. Then in September, Anthropic employee Jacob Coxon resigned on the grounds that the company’s researchers believed that it could kill us all by the end of the decade.
Three days later, Anthropic’s Chief Executive Officer Dario Amodei, released an essay that referred to both the huge benefits and the possible harms of AI, and called for slowing the pace of development so that we can develop the ability to control it. Other AI leaders, including Sam Altman, Demis Hassabis, and Elon Musk, voiced broad agreement with Amodei.
Trump was not having it. Instead, he called “AI doomerism” a “hoax” and proclaimed: “I am the hoax buster.” As with climate change, Trump put his own opinion above that of the acknowledged experts in the field, including Musk, whom he had previously described as a rare “super genius.”
‘Doomers’
The motivation for opposing any slowdown in AI development was made clear by another Department of “War” post, again with a photo of Trump, this time under the headline: “The new frontier of artificial intelligence belongs to us, not doomers.” Anthropic and Amodei then became the administration’s targets, perhaps because in February, Anthropic had refused to remove restrictions on the use of Claude, its world-leading AI model, to conduct mass surveillance of Americans and develop weapons that could be used without human involvement.
The first public suggestion that Amodei’s call for an AI slowdown was influenced by effective altruism appears to have been made on Sept. 15 by Undersecretary of Defense for Research and Engineering Emil Michael. In an interview on Fox News, Michael said the call for a slowdown “comes from a deep-rooted ideology that is pretty dystopian.” The department posted this section of the interview on X, under the heading: “Under Secretary Emil Michael on the dangerous ideology of effective altruism.” Michael did not explain why a movement that encourages altruism is “dystopian.” The news website Axios subsequently obtained a memo it said was “penned by a Trump political adviser” and “circulating within the White House.” The memo, headed “Effective Altruism: Opposition Research File,” describes me as “major intellectual precursor of EA” and “the philosophical edge of EA.” That exaggerates my significance. EA has many other influential philosophers among its adherents, including Toby Ord and Will MacAskill.
In typical “oppo” fashion, the memo rakes up some of my most controversial ethical remarks on issues that have nothing to do with EA, without indicating why I hold those views, or demonstrating that they have influenced EA. Other prominent EA figures are similarly subjected to irrelevant smears.
The administration’s attacks on EA show no awareness that the movement aims to save lives, reduce suffering, and prevent future catastrophes. A 2024 survey of approximately 1,850 people who identified as EAs indicated that 61 percent would allocate the movement’s resources to reducing the present suffering of humans and animals. Only 35 percent favored long-term causes, of which AI safety was then considered one. Since then, however, opinion within and outside the EA movement has become much more concerned about AI. An Ipsos poll released in March found that 66 percent of US adults agreed that AI development needs to slow down. The problem, of course, is how to achieve that on a global scale, given that no AI company, and no country, wants to be left behind.
I am not and never have been an “AI doomer.” I have often spoken out against what I regard as the excessive focus that some EAs have given to AI extinction risk. But controlling AI to reduce the risk of extremely serious consequences, even if they stop short of extinction of our species (Bill Gates said recently that AI could cause a billion deaths), is no longer a long-term issue. It is alarming that Trump is playing partisan politics with AI safety, as he did with climate change, thereby rendering impossible the global cooperation needed to reduce the risks they pose to our species’ well-being.
Peter Singer is professor emeritus of bioethics at Princeton University and visiting professor of medical ethics at the Center for Biomedical Ethics at the National University of Singapore. He is the author of “Animal Liberation, Practical Ethics, Ethics in the Real World, and The Life You Can Save” (Random House, 2009), and the founder of the nonprofit organization of the same name.
Copyright: Project Syndicate, 2026 www.project-syndicate.org



