Sex, drugs, and effective altruism: what EA and harm reduction can learn from each other

By Kristof Redei @ 2025-07-31T15:58 (+10)

This is a linkpost to https://www.moregoodlessharm.com/p/sex-drugs-and-effective-altruism

The main purpose of this Substack is to connect the communities of effective altruists (EA) and everyone else interested in evidence-based efforts to solve big global problems, with the world of harm reduction, which focuses on meeting people who take risks others don’t approve of where they are at, rather than attempting to coerce them into changing their habits. My feeling has always been that the two groups have things to learn from each other. This post sketches out some ideas on what those might be.

I’ll start by outlining some normative claims that touch on the philosophical underpinnings of both movements, what they share, and where they differ from each other. Then I’ll try to figure out whether harm reduction’s moral claims could or should appeal to people interested in EA’s approach to evidence-based ways of doing good at scale, and vice versa. This exercise provides a chance to look at what each school of thought has to offer the other.

Introduction

Effective altruism and harm reduction both start with the observation that the world contains vast, preventable suffering, much of which persists because dominant moral frames and policy defaults are poorly aimed. Each movement proposes a shift in perspective that makes previously invisible or neglected interventions suddenly obvious.

The pivot EA suggests is that people (especially in rich countries) with disposable resources can move from not donating, or donating based on sentiment and proximity, to deliberately maximizing impact, which most of them currently don’t do. A modest redirection of money, time, and talent guided by evidence and expected value can save or radically improve far more lives than intuition suggests. The reframe proposed by harm reduction is that policymakers, clinicians, and advocates can move from trying to coerce or shame people who take unpopular risks (psychoactive substance use, smoking, sex work) to empowering them by providing safer options, more accurate information, and non-judgmental services. Instead of trying to force abstinence from activities the majority rejects, harm can be reduced in the world as it is actually lived.

Both movements’ adherents claim that their signature interventions look radical only because our default moral lens is skewed. Shifting from charity-as-virtue to charity-as-optimization or from prohibition-as-protection to autonomy-plus-safety allows new possibilities to open. Beyond these similarities, however, there are also deep differences in philosophical foundations and political instincts.

Philosophical foundations

EA: from Mill and Singer to malaria and shrimp

While EA as a movement began in the late 2000’s, its intellectual genealogy goes back to philosopher Peter Singer, in particular his 1972 essay, “Famine, Affluence, and Morality,” and even farther back to a longer tradition of consequentialist and utilitarian thinkers like Jeremy Bentham and John Stuart Mill. Singer’s essay argues that if you can prevent something bad “without sacrificing anything of comparable moral importance,” you are morally obliged to do so. From this act-utilitarian perspective, outcomes are what matter, and the location of the people affected (whether they are in your hometown or on another continent), whether they are alive now or will be in the future, or whether they are even people or other sentient beings with moral worth, shouldn’t change the calculus.

EA then attempts to put these utilitarian ideas into practice in a pragmatic way. The Center for Effective Altruism defines four core principles of its ethos. Prioritization refers to the process of using evidence to rigorously weigh the relative impact of different actions we can take to help instead of relying on our intuitions. Impartiality echoes Singer’s imperative to weigh all those affected equally regardless of their distance from us in space and time. Open truthseeking refers to the habit of reflecting on our views regularly with an openness to change them as we discover new evidence. Collaborative spirit emphasizes collaborating with others on altruistic goals with “high standards of honesty, integrity, and compassion.”

The institutional face of EA, consisting of organizations like GiveWell (charity evaluation), Open Philanthropy (grantmaking), and 80,000 Hours (career advice), builds this into practice. To learn about what cause areas should be prioritized, they do things like estimate disability-adjusted life years (DALYs) saved by various interventions, support and learn from randomized controlled trials, weigh the moral significance of various species, and assign probabilities to various forms of AI risk. Even when EAs acknowledge “moral uncertainty” or allow pluralism, for example by allowing for virtue ethics or deontology as side-constraints in considering possible actions, the default goal is still maximizing expected value.

Harm reduction: rights-based consequentialist pragmatism

The harm reduction movement’s members come to it bringing with them a variety of ethical intuitions. Its roots are found not in abstract moral reasoning or an explicitly defined, shared philosophy and are instead practice-driven. It began as an umbrella term for a set of pragmatic responses to concrete problems encountered by practitioners working in a number of fields dealing with various forms of socially stigmatized risk-taking like sex work and intravenous opiate use. Nonetheless, it’s possible to identify a couple of conceptual strands that recur regularly in harm reduction discourse and drive its practice.

Harm reduction frequently centers a rights-based approach to morality. The National Harm Reduction Coalition in the US calls for respecting “the rights of people who use drugs” in its presentation of the principles of harm reduction and Harm Reduction International describes it as “grounded in social justice and human rights.” Philosophical pragmatism in the tradition Dewey, James, and Peirce has also been described as “one of the underlying theories of harm reduction” and lying “at the heart of harm reduction” as its practices grew out of on-the-ground experience rather than academic theorizing or top-down bureaucratic initiatives. Moral philosophers advocating for freedom as a foundational concept, like Mill’s articulation of the harm principle in On Liberty, were the first to make explicit the ideas that underpin the resistance to prohibitions on sex work and psychoactive substances. To thinkers like Mill, these prohibitions are harmful in principle, not because of their consequences, but simply because they restrict the individual’s liberty. Most closely connected to EA’s approach, utilitarianism plays a prominent role in harm reduction as it’s practiced, as the arguments for its superiority to criminal or medical framings of the problems it’s applied to tend to be justified on cost-benefit grounds.

How different are these approaches?

The approaches to ethics behind both EA and harm reduction contain a strong dose of consequentialism. For the former, this is an integral part of what distinguishes the framework from more conventional approaches to philanthropy. For the latter, consequentialism supplies just one of the multiple moral arguments for why it’s a better approach to addressing its areas of concern than competing models. Adherents of both regularly invoke straightforward cost-benefit analyses to convince stakeholders to adopt their approach. They differ in at least two important ways, though, with regard to additional moral commitments that affect what counts as a legitimate input into those analyses.

Harm reduction proponents center their ideas as a response to coercive and paternalistic measures. They may dismiss efforts that an EA might consider to net out positive if they involve what they see as unacceptable infringements on individual liberty to choose behaviors the majority regards as too risky or even irrational. In the stark words of psychiatrist Thomas Szasz, “freedom is more important than health.” For example, they may oppose, as a matter of principle, policies like tobacco taxation or framings of addiction as a disease to be cured that EAs would advocate for because a DALY-based cost-benefit analysis shows the possibility of a big win on purely utilitarian terms.

In turn, the principle of impartiality stressed by EAs is responsible for an important and distinctive aspect of their worldview. In an effort to take the welfare of moral persons from the far future into the same consideration as those currently living, they will seek out and incorporate information on the possible effects of interventions on future generations as much as possible. This is taken furthest through the idea, contested even within EA, of longtermism, the view that positively influencing the far future is the key moral priority of our time. This isn’t generally an instinct shared by harm reduction proponents, who tend primarily to be concerned with reducing the immediate suffering of those taking unpopular risks. Because some of the most frequent criticisms of harm reduction approaches to psychoactive use, for example — like the justification of prohibition of some substances alleged to be “gateways” to more dangerous ones, or the idea that the prevention of youth uptake needs to be a strong consideration when promoting less risky forms of use — appeal to long-run considerations, they are likely to view such arguments with a more skeptical eye than most.

In practice, both traditions are more pluralistic than a strict reading of their slogans might suggest. Many EAs appeal to moral uncertainty and accept side‑constraints like autonomy and consent as genuine moral reasons that can limit policies with otherwise high expected value. Some also draw on virtue‑ethical considerations about integrity, trust, and movement‑building effects. Conversely, harm reduction advocates routinely engage on consequentialist terms: they run cohort studies and natural‑experiment analyses, contest “gateway” claims with longitudinal evidence, and argue that syringe access, naloxone, safe consumption sites, or low‑risk nicotine products outperform punitive or abstinence‑only approaches on DALY-convertible measures like infections, overdoses, and mortality. There is internal diversity of thought on hard cases like differential tobacco taxation or youth‑access restrictions where some harm reduction proponents accept least‑restrictive nudges to shift behavior, while others view these as unacceptable paternalism. EAs aren’t averse to weighting autonomy costs and legitimacy effects in their analyses, and harm reduction organizations regularly quantify outcomes to prioritize campaigns and secure scale. So the disagreement in specific cases is usually less about whether consequences matter and more about which constraints are to be included and how heavily they should be weighted in the moral calculus.

What can EA learn from harm reduction?

EA and harm reduction share a taste for pragmatism, but the latter’s history inside adversarial, morally charged policy domains surfaces blind spots that EAs are prone to miss. None of these are totally unfamiliar or undiscovered, as they’ve been raised as critiques of EA in the past, but harm reduction provides both a vivid set of examples and a number of tested practices to help counter them.

One of EA’s strengths is the relentless focus on quantification above anecdote and emotional salience. But many of the most consequential harms in harm reduction domains are relational and institutional, and therefore rarely map cleanly to DALYs. Loss of trust in authorities, fear of police contact that keeps people away from services, and criminal records that foreclose jobs and housing are all factors that undeniably affect the total cost of an intervention, but analysts may be tempted to treat them as second‑order. Harm reduction’s rights language functions as a guardrail in encouraging a perspective that sees coercion and degradation themselves as public health risks. Policies that look efficient on paper can unravel when people avoid them or when the legitimacy costs poison future cooperation. A harm reduction-informed perspective can encourage EA evaluations to consider when and how impacts of an intervention on autonomy, dignity, and institutional trust can be made first-class considerations in modeling them, rather than more implicit caveats.

Unintended consequences are at the core of why harm reduction exists as a concept. Drug and alcohol control offer a running lesson in perverse effects, as when supply‑side crackdowns lead to more potent and adulterated products or bans on sterile equipment drive blood‑borne infections. Tobacco regulation has its own versions in the form of flavor bans nudging adults back to cigarettes or into illicit markets and advertising prohibitions that incidentally suppress truthful risk communication. Because of this history, harm reduction’s habit is to ask, before acting, how might this backfire, and on whom? For EAs, this highlights the importance of considering backfire scenarios in cause area evaluations. If an intervention expands a black market, erodes service uptake, or generates political backlash that crowds out better policies, our models are improved when those pathways receive explicitly quantified weight.

For EAs, especially those that are strongly convinced by longtermist reasoning, future people matter just as much as those alive today. But harm reduction experience shows how less determinate future worries, like “gateway” hypotheses or speculative effects on normalization of substance use can be used to block concrete, tractable benefits for living people. The corrective is not necessarily to ignore the future, but to discipline it by requiring explicit estimates before letting a conjectured risk outweigh immediate, well‑evidenced gains. In practice, this pushes toward reversible policies such as piloting and monitoring safe‑consumption sites, permitting low‑risk nicotine products while tracking youth uptake, and sunsetting coercive elements unless supported by data. This can ensure that present benefits are realized while still bounding future risks.

One of the most well-known slogans of harm reduction advocates, adopted from the disability rights movement, “nothing about us without us,” can also serve as a useful reminder for EAs. People who use illegal substances, smoke, or sell sex possess local knowledge about incentives, enforcement patterns, and behavioral substitution that formal models can easily miss. Their participation improves domain knowledge, helping to reveal bottlenecks and surface overlooked harms and opportunities early. Prioritizing affected‑community co‑design could improve EA projects across cause areas. This includes things like budgeting for peer researchers, sharing intermediate findings for critique, and planning for legitimacy effects (compliance, uptake, political durability) as part of the expected value calculation.

Taken together, these lessons point to a more rights‑sensitive EA practice: quantify relentlessly, but price autonomy and trust; model backfires, not just benefits; resist letting hazy long‑run fears veto solid near‑term gains; and upgrade participation from courtesy to core methodology.

What can harm reduction learn from EA?

EA’s distinctive contribution is methodological: make trade‑offs explicit, quantify expected value, publish assumptions, and update in public. Harm reduction campaigns often have compelling narratives and strong prima facie evidence, but budgets are finite and opportunity costs are real. Systematically comparing options with transparent cost‑effectiveness models rather than relying on intuition or political tractability alone would help harm reduction allocate marginal dollars between, say, naloxone distribution, syringe services, contingency management for stimulants, smoking‑cessation through low‑risk nicotine, or housing‑linked supports. EA orgs like GiveWell make their spreadsheets, moral weights, and sensitivity analyses public. They routinely find that some programs are an order of magnitude more cost‑effective than others, and they communicate that clearly (e.g., saving a life for on the order of $3,500–$5,500 via top global‑health programs). Adopting that level of quantitative transparency, and inviting critique, could improve harm reduction credibility with funders while disciplining internal priorities.

Relatedly, EA’s willingness to privilege measured impact over empathetic pull offers a useful corrective. It is often more emotionally satisfying to fund broad “wrap‑around” services than narrowly targeted, high‑leverage interventions. Yet the evidence that certain harm reduction staples avert deaths and infections at very low cost is strong: community naloxone distribution prevents overdoses and is cost‑effective with modeled estimates of one death averted per roughly 89 kits distributed; syringe services halve HIV/HCV incidence and, combined with medication, cut transmission by more than two‑thirds. Building comparable cost-effectiveness analyses around such effect sizes, and letting the numbers guide marginal spending, can mean backing programs that feel less comprehensive but save more lives.

EA also stresses prospective risk analysis: model backfires, specify priors, and test sensitivity to pessimistic assumptions. Harm reduction’s history teaches that blunt controls can backfire, but the mirror image is also true: some structured constraints appear to reduce harm at population scale. Scotland’s minimum unit pricing for alcohol, a paternalistic nudge many harm reduction advocates distrust, was associated with a 13.4% reduction in alcohol‑attributable deaths and significant hospitalization declines, especially in deprived groups. Tobacco taxation shows similarly robust links to lower prevalence and even downstream outcomes like reduced infant mortality. Harm reduction can incorporate these findings without abandoning its focus on rights by demanding high‑quality evidence of benefit, preferring least‑restrictive designs, and quantifying regressivity, illicit‑market displacement, and enforcement harms.

Finally, EA’s norm of “open truth‑seeking” is operationalized through preregistration, randomized controlled trials and natural‑experiment designs, and adversarial collaboration. Harm reduction practice already uses many of these tools, but making them movement norms by publishing forecasting intervals for expected deaths averted and preregistering before‑after evaluations of policy changes would strengthen advocacy and guard against motivated reasoning. Where constraints clearly fail, harm reduction should emphasize it, such as COVID‑era evidence that relaxing methadone take‑home rules improved retention and client experience. Where there’s robust evidence of constraints having a strong net benefit, harm reduction can consider endorsing them as provisional, monitored, least‑restrictive instruments in a broader rights‑respecting framework.

Harm reduction can borrow some of EA’s best habits of publishing models, comparing marginal cost‑effectiveness, and explicitly pricing in dignity, regressivity, and enforcement harms. That combination keeps harm reduction anchored to its rights‑based ethos while maximizing the lives and suffering averted per dollar.