Why Effective Altruism: A Response to Wu

27 mins read
Image provided by Koji Flynn-Do

Flynn-Do is co-founder and co-president of the Swarthmore Effective Altruism Society.

In a recent opinions piece for The Phoenix, Megan Wu argues that Effective Altruism (EA) — a social and intellectual movement that advocates the use of careful reasoning to determine the most effective ways of improving the world, and then acting on those — fails to live up to its name. Firstly, I’d like to thank Wu for her thoughtfully considered writing on EA. It is easy to not take ideas seriously and to instead throw stones without making a real effort to understand what it is that one is criticizing, and Wu skillfully avoids doing so. The two of us have spoken in the past at some length about these questions, and to see her views  written out clearly is illuminating. 

At the outset, let me be clear that I agree with Wu on many issues. And so, I must say, “Not Effective and Not Altruistic” was a surprising subtitle for an article in which she writes, among other things:

“Effective Altruists can certainly do, and have done, a lot of good.

I don’t mean to discourage doing good, so let me be clear that being an Effective Altruist is better than being amoral or evil, and the actions of Effective Altruists have done tremendous good.”

In this piece, I lay out the case for Effective Altruism, as I see it. Then, I respond to Wu’s objections in turn.

What is the case for effective altruism?

Wu writes that she “understands the motivation behind Effective Altruism all too well,” as “a lifestyle of levelheaded thinking and ethical action which shields both the doer and recipient of altruism from the vicissitudes of the modern world.” Though people’s reasons for getting involved in Effective Altruism are no doubt multifaceted, this certainly doesn’t reflect what attracts me to EA. 

Taking seriously a commitment to helping others as much as possible and addressing the world’s most pressing injustices entails grappling head-on with the many ways the world is horribly broken; it means looking at the horrors of suffering, systemic inequality and oppression, and existential dangers without flinching, and resolving to do what we can to abolish them. I think it is absolutely criminal that the world remains replete with such suffering and deprivation while the rich get richer, and I think we can do things to change that.

What kinds of things might those be? I take some of the core tenets of EA to be:

Radical empathy. One way we might fail to contribute as much as possible to solving pressing injustices is by failing to even recognize the moral significance of others. Historically, it has always been the case that entire populations have been dismissed, oppressed, and treated as lacking in fundamental dignity and moral worth. People today tend to be biased, for example, to others of their same nationality, even when the effects of our actions have enormous consequences globally. Many in EA, therefore, seek to widen their circle of moral concern to all those who ought to be included, irrespective of where they happened to be born, their gender and sexuality, and sometimes even when they happened to be born or what species they are. Then, people committed to EA seek to live in accordance with those values.

Taking a long-term perspective. EA tends to be “scope-sensitive:” it cares about how many people are suffering or facing injustice. This often means caring for the interests of those who will follow us, especially when there will be even more of them. For those reasons, many people in EA care a lot about promoting justice, well-being, and democracy in the very long-run — not just today.

Effective career choice. After graduating and entering the workforce, we will spend roughly half of our waking hours on our careers. If you can find a career trajectory that’s 100 times higher impact than your current default, then you can do in ten years what would otherwise have taken people like you 1,000 years — and certain approaches to one’s career offers exceptional opportunities to make the world a better place. At Swarthmore, the Effective Altruism Society is especially interested in disrupting the finance, consulting, and tech pipeline that exists at many top colleges. 

Thinking on the margin. Some problems, and hence ways of helping others, are much more neglected than others. One can therefore often have an outsized impact by focusing on such issues. The fiftieth person working on some specific cause is likely to be able to make more progress than the ten-millionth, simply because — by that point — the best opportunities (“low-hanging fruit”) have been taken. It’s often the case that pressing problems don’t receive much attention. “Neglected cause areas” are neglected precisely because they don’t get attention from for-profit companies or because philanthropists are looking to bolster their own reputation by focusing on trendy causes.

I think these are compatible (and often harmonious) with many different ethical and political theories. Indeed, members of Swarthmore EA also are and have been involved with DSA, Sunrise, and other social movements precisely because of these principles. I’d also add that there is a kind of elite philanthropy that is less about promoting justice and more about reputation laundering, propping up existing systems of oppression, and covering up exploitation. EA was very much born out of a reaction against this damaging and fraudulent system.

On the basis of these views and others, those interested in EA seek the best ways to contribute to the common good and then try to do just that. I’m most inspired by Swarthmore alums who, motivated by EA ideas, have gone on to work directly things like preventing the next pandemic (Jaime Yassif, senior director and lead scientist for the biosecurity program at NTI), identifying the most effective ways to help those in extreme poverty (Natalie Crispin, vice president at GiveWell), and guiding the development of artificial intelligence toward trustworthy, robustly shared and beneficial outcomes for humanity (JJ Balisanyuka-Smith, engineer at Cohere).

With that established, let’s turn to Wu’s critiques.

Are “conditional” efforts to improve the world altruistic?

Wu argues that, because Effective Altruism exhorts us to pick the most effective means of helping others with the resources we devote to that purpose, it is thereby not altruistic. “Effective Altruism is not actually altruistic because the purportedly altruistic choices the movement prescribes — donations, career choices, etc. — are all done conditionally … [on] some expected payoff.” That is, by selecting where to devote one’s energy on the basis of what one thinks will help others the most, one is no longer acting altruistically.

This is remarkable criticism to level against a movement which advocates that we should seek to help others as much as possible, often at considerable personal sacrifice — changing careers, giving large fractions of one’s income, and trying to maximally channel one’s efforts to help others into the most effective methods. Wu wants to call this something other than altruism; it is a curious parlance.

The deeper problem with this line of reasoning is that it implicitly imagines a world in which we do not have to choose how we allocate our efforts. But the world is broken in so many ways; by working on one thing, one is thereby making the choice to not work on everything else. We have roughly 80,000 hours in our careers. This is a lot of time, but it’s not infinite. It means we must choose which problems to devote our effort toward. Of course, many problems are intersecting in their very nature: injustice in one area begets and compounds injustice in another. But denying that we must choose will only mean we make worse decisions. Failing to think explicitly about such choices means that more suffering continues unabated, that more injustice is allowed to continue, that more potential for flourishing is lost. This is a central fact that is important to internalize: the world has many problems, and while we can maintain solidarity with victims of injustice everywhere, we cannot but choose which efforts to focus on in our own lives. 

If Wu wishes to call these efforts to improve the world non-altruistic, that’s fine. But the material fact remains that people in EA are trying as much as possible to help others. That is the point.

The two thought experiments Wu introduces might be more troubling than the linguistic game discussed above, and I wish to address these head-on. In the first, she suggests that because she is deaf, given “a scenario where only me or my hearing twin can survive,” Effective Altruism would choose the hearing twin. It’s surprising that this is leveled as a criticism of a movement that is in large part dedicated to improving the health and wellbeing of the worst off in the world. The actual choice presented to us by the world is nothing like the hypothetical posed. The choice is not between saving one and condemning another. Instead, it is much starker: do we think it is better to spend our money on luxury goods or on protecting hundreds of people from malaria, which still afflicts millions yearly? Is it legitimate to choose to transfer cash directly to the world’s poorest people rather than unconditionally assist whomever is in our vicinity? To me, the choice is clear.

In the second thought experiment, Wu claims that Effective Altruism commits one, given the choice between saving a child or a Picasso painting from a burning building, to saving and selling the painting and then donating the proceeds. She claims that, following the logic of Effective Altruism, one must either bite the bullet and save the painting or, in order to justify saving the child, speculate that the child’s future contributions to society will end up helping more in the long run.

Are these really our only choices? They are not. I, for one, would save the child. I do not wish to live in a society that values profit over people, nor one that views people as means toward ends, nor one in which social trust is so low that one would even hesitate before saving the child. It is clear to me that actions which encourage such a society are incredibly damaging when considering all their effects. Seeking to help others as much as possible does not commit one to damning the child. Indeed, a more thorough accounting of harms and benefits reveals that, in the world we encounter, it is essentially never a good idea to act in contravention of such straightforward moral principles. Effective Altruism says no different.

Is Effective Altruism just about maximizing quality-adjusted life-years?

Wu believes that “Effective Altruism uses quality-adjusted-life-years, or QALYs, as a metric for quantifying good.” This view is mistaken. Contra her claims to efforts at charitability, Wu has set up a straw-person to knock down.

First, the supposed conception of moral value as simply maximizing QALYs is comically narrow. QALYs are a measure of health, and not even the most bullet-biting utilitarian believes that health is all that matters. We can all be in agreement that a good life is not reducible to mere health and involves a commitment to justice, flourishing, and self-actualization.

The much more reasonable view is that when focusing on efforts targeted at improving human health, we ought to prefer actions which improve human health more than those which do less. This may sound banal, but is in fact crucial. The most effective ways to help others with a set of resources within a given issue often differ by orders of magnitude. For example, the most cost-effective of common interventions to prevent and treat HIV/AIDS is estimated at 1,400 times as cost-effective as the least. One would not know this without investigating the numbers.

Such enormous differences are matters of life and death. As philosopher Toby Ord writes, “[i]n practical terms, this can mean hundreds, thousands, or millions of additional deaths due to a failure to prioritize. In non-life-saving contexts it means thousands or millions of people with untreated disabling conditions.”

GiveWell, as Wu notes, is one organization devoted to the project of identifying the most cost-effective ways to improve human health in contexts of extreme poverty. She objects, though, to its spreadsheet for moral weights as being “devoid of units for the values listed and contain[ing] no rationale for the numbers” assigned to various moral weights of, for example, saving the life of a child under five from malaria compared against doubling income through direct cash transfers. I commend Wu’s research into GiveWell’s methodology, but she has stopped a few steps too early.

From where does GiveWell derive moral weights for different outcomes? In large part, these weights come from polling the likely beneficiaries of programs in extremely poor regions of Ghana and Kenya, asking them how much they value saving lives of different ages compared to lifting people out of poverty through direct cash transfers. That this is unprecedented (as far as I know) speaks to the fruits of a rigorous approach to helping others and to the fact that EA organizations take very seriously the idea of listening to the historically marginalized and ignored communities that it aims to support. 

This stands in sharp contrast to Wu’s claim (oft-leveled, but seldom justified) that “Effective Altruism’s viewpoint towards the developing world is that of western saviorship.” A more accurate characterization of EA support for specific health and poverty interventions is that of a mass transfer of wealth directly from countries which have unjustly acquired it to the most effective ways to aid those most harmed. Indeed, organizations recommended by GiveWell like Evidence Action and GiveDirectly are headed up and staffed overwhelmingly by people from the communities and countries they serve. As the post “How to not be a ‘white in shining armor’” from the GiveWell blog makes clear: “We fundamentally believe that progress on most problems must be locally driven. So we seek to improve people’s abilities to make progress … and thus maximally empower people to make locally-driven progress on other fronts.”

There is also great virtue in the numerical approach that EAs often endorse. It is much harder to accidentally deceive oneself about the effect of one’s actions when we make serious efforts to rigorously compare the options available to us, be it in our careers or in evaluations of various programs. And, when certain tactics are paying off or exceeding expectations — as in the case of campaigns against corporate abuse of animals on factory farms — one can correspondingly devote more effort to those.

I hope it is clear that I do not mean to say that the only things that matter are those which are easily quantifiable. I am in agreement that if one only considers policies that benefit the world in obviously quantifiable ways, one will miss out on structural improvements to make the world better and more democratic. Still, one can identify indicators that correlate with the terminal goals of interest, like union density in an industry, or voter turnout in historically marginalized communities, and so on. This is the approach taken by EA-inspired organizations like Just Impact, which works to end mass incarceration in the U.S. by directly supporting transformative leaders with a track record of success in preventing jail and prison construction, including and especially formerly incarcerated people.

Indeed, the account of EA-as-QALY-maximization struggles mightily to grapple with actually existing EA actions, careers, and priorities. Consider as one example EA’s longstanding interest in preventing pandemics, particularly globally catastrophic ones. Was this priority arrived at by QALY-crunching and deciding that working on this problem yields more QALYs? No. EAs work on preventing pandemics because the historical record shows that they are incredibly dangerous, far more likely than people think (and hence underinvested in), and there is ample opportunity to avert the incidence of pandemics and bioweapons that is incredibly neglected. Other pressing problems like runaway extreme climate change or the mass suffering of billions of animals in concentrated animal feeding operations similarly do not rely on QALY-crunching.

Ethical Paradigms, Practical Significance?

Despite her repeated gestures to the commonsense meaning of words, or the “average person’s views of moral goodness,” and her claim that “Effective Altruism smuggles in a niche definition of goodness,” Wu in fact advocates for a very specific branch of moral philosophy called virtue ethics. On this account, ethics is about the cultivation of virtue in oneself, seeking to improve oneself in wisdom, courage, and so on. To my mind, this view is irredeemably self-centered.

I am personally far more attracted to a view of ethics which is other-centered. That is, rather than my own personal preferences and goals, I think that the flourishing, suffering, and (in)justice that others experience is what is most morally relevant.

My impression is that this philosophical difference motivates Wu’s objections to Effective Altruism, which she views as fundamentally and intrinsically utilitarian. But no commitments to esoteric paradigms in moral philosophy are required to endorse seeking to improve the world as much as we can, ameliorating unnecessary suffering and injustice.

We should all be able to agree that it is a tragedy of unthinkable proportions that fifteen thousand children die every single day of easily preventable causes, or that climate catastrophes will kill hundreds of thousands and displace millions more in the coming centuries, or that billions of male chicks are thrown alive into grinders yearly because they cannot lay eggs. And we should all be able to agree that, to the extent that we can use our careers, our time, and our money to make a big difference on these pressing problems and others, doing so would be a good thing.

So it would be helpful to see where the proverbial rubber hits the road. To the extent that these criticisms are largely theoretical disagreements over which paradigm one ought to adopt in moral philosophy that do not bear upon our really existing choices, I think that their significance is much diminished in actual practice. Indeed, Wu writes:

“In critiquing Effective Altruism, I do not intend to argue here that Effective Altruism’s preferred causes are bad ones, nor that any particular action which Effective Altruism recommends is harmful.”

In writing this response, I have no intention to close the book on the merits of the approaches, ideas, and actions suggested by EA thinking. In particular, there remains so much to learn about the most effective ways to tackle the world’s most pressing injustices and the problems on which one might be able to make the most difference. What Wu views as indicative of self-serving interest — “more research into what Effective Altruism’s priorities should be” — I view as a sign of intellectual humility and a recognition that the situation we face today in the world is enormous, complex, and resists easy comprehension.

Let me end on points where, I believe, Wu and I agree. We agree that moral goodness is multidimensional and not easily reducible. We agree that many terrible injustices continue, that we can make real progress in addressing those, and as such we ought to “make conscientious choices and try to do well by others, both in our career trajectories and everyday lives, while still acknowledging that we are a part of a vast and complicated world.” And we agree that efforts by those inspired by EA have, in her words, “done tremendous good.”

I thank Wu for the opportunity to discuss these questions in an open forum, and I hope we can continue to find common ground.

Edward Tranter read an early draft of this article and provided feedback. He is co-president of the Swarthmore Effective Altruism Society.


  1. I appreciate the chance to read your thoughts, Koji!

    The claim that “Wu in fact advocates for a very specific branch of moral philosophy called virtue ethics” is factually incorrect. I did reference a distinction between thick and thin concepts in my piece, which was originally proposed by virtue ethicists, but using this distinction in support of my argument obviously does not amount to advocating for their views. There have been many defenses of virtue ethics written by advocates across countless historical eras and religious traditions. Unlike these, my piece should not be taken as advocacy for virtue ethics, even though I am grateful to be as influenced by their thinking as we all are.

  2. dawg I hate to say it but it feels like you didn’t read the other article and took it a bit personally

  3. dawg I hate to say it but I think you didn’t really get the other article and took it a bit personally

  4. Let’s begin with the following:
    One of EA’s core ideas is taking a rigorous analytical approach to maximizing benefit to others. Part of this rigorous analytical approach involves calculating the expected value of our actions in terms of well-being Remember, expected value = probability * magnitude, and in this case the unit of magnitude is quality adjusted life years, or if one is opposed to using such “narrow” measurement units, general well-being.
    Regarding the Picasso argument:
    The thought experiment discussed by Ms. Wu indicates that given the choice of saving 1 child vs saving the Picasso, which could be sold for a sufficient amount of money that would fund saving hundreds or thousands of children, the effective altruist is necessarily committed to saving the Picasso lest he be inconsistent with EA principles.

    Mr. Flynn-Do’s response is that he would save the child, because actions that “encourage” a society in which “profit[s] over people” are valued, “people [are treated as] as means to ends”, or “low social trust” exists are contrary to EA principles. Thus, consistency with EA, according to Mr. Flynn-Do, requires saving the child rather than the Picasso, nullifying the argument of the thought experiment.
    The author here appears to construct an implicit standard that actions that “encourage” a society with poor moral values are at least as bad as the decision to avoid saving hundreds or thousands of people.

    Notwithstanding the lack of any specificity as to what constitutes “encouragement” of poor moral values in the first place, this kind of general disapprobation does not compare to the specific large benefits generated (in terms of QALYs/well-being) by saving the Picasso and using the proceeds generated from its sale to save hundreds of children. To believe it does compare favorably, one has to make unfounded assumptions about the social visibility of one’s actions in this specific scenario and its knock-on effects/influence on the choices of others. This is the epitome of “fuzzy math” and contravenes the explicit principle and implicit ethos of rigorous charity that suffuses the EA movement. The author asks us to engage in precisely the sort of prediction gymnastics that is used by, for example, traditional philanthropists who justify lavish donations to already wealthy cultural institutions on the grounds of the ostensible intangible cultural benefits created by their work (eg the Sackler family’s lavish donations to museums).

    A few more points:
    1. The author’s argument that saving the Picasso rather than the child would implicitly value “profits over people” is missing a premise, and seems dubious upon further inspection. In the hypothetical, the altruist is forced to save the Picasso rather than the child precisely because saving the Picasso would result in MORE PEOPLE being saved, not because of any intrinsic value to the cash generated by saving and selling the Picasso. The author never explains why he believes saving the Picasso would implicitly value profits over people.

    2. Treating people as “means to ends” – first, if treating people as means to ends is bad, we may as well do away with EA’s instruction to people in wealthy countries to live their lives as instruments – means – to creating benefits to others. EA necessarily commits itself to treating individuals as means to the end of maximizing well-being for others. The author’s response to the thought experiment here is internally contradictory with EA principles. Second, the argument is again missing a premise – how does saving the Picasso rather than the child treat the child as a means to an end? In this thought experiment, the child is not used as a means to accomplish any particular end – rather, the effective altruist simply decides to save the Picasso. Consider the counterfactual: if the child was not there in the burning building, the effective altruist would be obligated to take the exact same action – save the Picasso and sell it to generate cash that can fund efforts to maximize quality adjusted life years/general well-being. Contrast this with the common variation of the trolley problem that involves pushing the large man in front of the train to save 5 other people – in that case, the counterfactual scenario (if the large man was not present in the scenario), prevents us from taking any action, since the large man’s presence is necessary as a means to the end of saving the other 5 people. (See here – https://en.wikipedia.org/wiki/Trolley_problem#The_large_man – for context if you are unfamiliar with the scenario discussed).

    3. Social trust – this is, again, a fuzzy concept used to wave away a much clearer expected value calculation. Moreover, it’s not clear that failing to save the child would constitute a breach of social trust – if it does send any sort of message to the wider “society”, it’s that the individual has chosen to maximize well-being, which an Effective Altruist would endorse. The author here continues to use nebulous knock-on social/cultural effects to distort the conclusions of EA. It’s just as plausible that the knock-on social effects of selling the Picasso and using it to maximize well-being is enormously beneficial to “social trust.”

    Regarding the other argument proposed by Ms. Wu, that “Given a scenario where only me [deaf] or my hearing twin can survive, the Effective Altruist will say that it is a clear choice which one of us should live.”

    The author’s response is simply the following: “It’s surprising that this is leveled as a criticism of a movement that is in large part dedicated to improving the health and wellbeing of the worst off in the world. The actual choice presented to us by the world is nothing like the hypothetical posed.”

    Even if it’s true that the current circumstances of the world do not give effective altruists the choice presented by the thought experiment, this is hardly a sufficient defense of EA’s underlying principles. The author’s response to this thought experiment, proposed to illustrate the repugnant conclusions mandated by EA’s principles, is effectively “well, that wouldn’t happen in the ‘actual choice[s]’ presented by the current world.”

    This is poor reasoning if we care about acting on ethical principles that reflect our moral intuitions in any meaningful sense. Moreover, it stymies allegedly ‘real-world reasoning’ as well – what does the effective altruist do, if, when working to benefit impoverished people, they are actually forced to make this choice? EA’s principles require them to follow a decision process that is intuitively morally repugnant to the vast majority of us. The author, meanwhile, has no advice to the effective altruist facing this decision because he declines to seriously consider the scenario.

    The interplay between the author’s and Ms. Wu’s ideas is illuminating in that it demonstrates the intellectual bankruptcy of the EA movement. Its members likely have good intentions, and it’s certainly possible that individual effective altruists happen to have net positive actions, morally. If this is the case, however, this will not be because of the EA movement’s principles and directives – it will be IN SPITE of, or unrelated, to them. It’s a bit like homeopathic remedies – they have no effect at best, and may actually hurt someone. Nonetheless, it’s entirely possible that a well-meaning caregiver administers the homeopathic remedy and the patient gets better. If the patient does heal, however, it will not be a result of the homeopathic remedy.

    EA’s good vibes and gestures toward charity are admirable, but vibes and ethos do not substitute for sensible moral principles. Perhaps the EA movement should spend some time educating themselves on moral philosophy in addition to their PR campaigns.

    brackets in quoted phrases are my own additions for context and clarity

    • Imagine writing this about a movement started by Oxford moral philosophers :/

      More seriously, I think it’s important to remember that this is an article for a general audience. I’m pretty sure Effective Altruists value having rigorous philosophical arguments very highly.

      I feel like both the critique and defense can be synthesized into: EA does a large amount of measurable good, but doesn’t solve everything. This is obviously the case, but saying that would not drive traffic like Megan’s clickbait title.

      • If EAs love rigorous philosophical arguments, where were they in this article?

        Like the original commenter said, all I see is two mishandled thought experiments, contradictory claims about having no moral framework while simultaneously presenting EA as a movement with choice-guiding principles, and misleading treatments of things like Just Impact, whose advocates, from a scope-sensitive and “thinking at the margins” perspective, are not acting in line with EA priorities.

        Plus some vibes!

  5. While I do think that Wu’s point about how Effective Altruists conflate “desirability with moral value” makes a lot of sense, I also agree with this article that there are a lot of definitions and semantics going on in her article. After all, what defines moral value? She offers no other true alternative outside virtue ethics, which is purely her opinion. What if someone personally finds it more moral to save more lives and so saves the Picasso? You could argue that is “amoral”, but based on what definition of moral?? Ultimately, it just comes down to one person’s opinion about what is moral and what is not. After reading both of these articles (which are so well written, by the way, so thank you to both authors for your research and politeness), and a couple of opinions in the replies, I think that we are all at a level where a lot of the disagreement with EA is purely opinion based about what is “morally good” etc. That’s pretty satisfying for me. Thanks, both of you! Now I don’t have to go read up on EA by myself, I pretty much understand it all haha.

Leave a Reply

Your email address will not be published.

The Phoenix