← Back to feed

The Economist published a cover essay on October 1 arguing that effective altruism has become “this century’s biggest idea,” reaching from Oxford philosophy departments into corridors of institutional power with what it called remarkable speed. The claim is not obviously wrong. What the essay underplays is the mechanism: EA did not spread because it convinced donors to give more efficiently to bed net charities. It spread because it staked out a position on transformative AI a decade before the rest of the world decided transformative AI was real.

What Effective Altruism Actually Is

Effective altruism, as a formal intellectual movement, emerged from Oxford in the late 2000s. Its central claim is that the moral imperative to do good should be pursued with the same rigour applied to empirical questions: measure outcomes, compare interventions, allocate accordingly. Early EA focused on global health and development, where randomised controlled trials could rank charities by cost-effectiveness. Global poverty was tractable. Mosquito nets were cheap.

That version of EA attracted a specific early audience: quantitatively inclined professionals who found moral philosophy off-putting but found spreadsheets compelling. The movement’s early institutional form, GiveWell, published data. Peter Singer’s philosophy provided the framework. William MacAskill’s 2015 book Doing Good Better gave it mass market packaging.

The movement’s centre of gravity shifted when EA thinkers applied the same framework to long-term existential risk. If the number of future people is very large, and if advanced AI could either dramatically extend or dramatically shorten the expected number of future lives, then AI safety becomes by far the highest-priority area for EA resources. That argument appeared in EA circles roughly a decade before the mainstream tech press treated artificial general intelligence as a near-term planning concern.

The Tech Money Followed the Logic

EA’s institutional footprint in AI is substantial and predates the current boom. The Machine Intelligence Research Institute, founded in 2000, was an early EA-affiliated organisation focused on AI alignment. Open Philanthropy, funded primarily by Dustin Moskovitz and Cari Tuna, became the largest single funder of AI safety research globally, backing the Center for AI Safety, ARC Evals, and dozens of academic alignment groups. Dustin Moskovitz also made personal equity investments in Anthropic, separate from Open Philanthropy’s grant-making activity.

Anthropic, valued at $61.5 billion after its March 2025 Series E, carries that EA-adjacent network in its cap table alongside the philosophical background of its founders. Dario Amodei has publicly distanced himself from the EA label — in his telling, the company’s safety focus is empirical rather than ideological — but the institutional and personal networks between Anthropic leadership and the EA ecosystem are documented and close.

OpenAI’s founding circle also included figures with EA adjacent views, and its own safety board drama in 2023 played out partly along lines that tracked internal debates about long-term AI risk.

From Philosophy to Regulatory Inbox

The more consequential shift is from philanthropy to policy. EA-affiliated researchers and organisations have testified before the US Senate and House. The Center for AI Safety’s 2023 statement on AI existential risk — signed by Geoffrey Hinton, Yoshua Bengio, Sam Altman, and Demis Hassabis, among others — drew directly on frameworks developed in EA-linked research. EU AI Act deliberations included substantial input from safety-focused research organisations that operate within the EA orbit.

The AI Safety Institute, established by the UK government in late 2023, was staffed heavily from a pool of researchers trained at EA-adjacent institutions. Its US equivalent, the AI Safety Institute at NIST, has faced criticism from AI capabilities advocates who argue that safety-focused oversight is slowing frontier development.

That tension — EA-influenced safety advocates pushing for precaution, AI capabilities advocates pushing for speed — is now the central conflict in AI policy in every major jurisdiction. The Economist is correct that EA has reached the corridors of power. It reached them specifically because the policy debate over AI risk is, in structural terms, the same debate EA has been rehearsing internally for fifteen years.

The Pushback Is Real

The criticisms of EA’s influence over AI policy are not trivial. Opponents argue that the long-termist framework that makes AI existential risk so central to EA priorities is philosophically contested and empirically unfalsifiable. Researchers who study near-term AI harms — bias, misinformation, labour displacement, surveillance — argue that EA’s institutional funding flows have structurally under-resourced those problems relative to speculative long-run scenarios.

The FTX collapse in 2022 dealt EA reputational damage it has not fully recovered from. Sam Bankman-Fried was EA’s most prominent public face and largest donor; his conviction on fraud charges created a specific problem for a movement whose brand rested on empirical rigour and ethical consistency.

And EA’s route to Washington runs through the same tech billionaires whose companies are building the AI systems regulators are trying to evaluate. That creates a structural conflict: the funding that sustains EA-aligned AI safety research also funds the organisations whose outputs that research is meant to scrutinise.

The Shape of What Is Coming

The Economist’s framing — EA as an idea that “conquered the world” — overstates the consensus. EA has filled specific positions in the AI policy debate, in specific jurisdictions, with specific funding streams. It has not won the argument. What it has done is arrive first, with a framework already developed, when the world decided the argument was urgent.

Whether that first-mover position translates into durable policy influence depends on questions EA cannot control: whether frontier AI produces the catastrophic risks that justify the precautionary stance, whether the near-term harms that critics prioritise become the dominant policy concern, and whether the movement’s institutional credibility survives further scrutiny of the people and money behind it.

The idea is influential. Whether it is correct is a different question, and one that governments are now being asked to decide before the answer is knowable.