Effective altruism

One of the endless series of panel discussions on EA, this one discussing AI and existential risk. The leader of effective assholism^[1] is in the center. Pictured, left to right: a moderator, Nick Bostrom, Elon Musk, Nick Soares, and Stuart J. Russell

Thinking hardly or hardly thinking? Philosophy

Major trains of thought

Logic Morality Reality

The good, the bad, and the brain fart

Sin Mind Monism Bushido Secular humanism Non-Overlapping Magisteria Euthanasia W.V.O. Quine

Come to think of it

Religion Science Philosophy of science Ethics Psychology v - t - e

Effective altruism (EA), and its umbrella concept longtermism, is a quasi-utilitarian^[3] movement to change the world through making carefully-targeted charitable donations — not only through making carefully-targeted charitable donations, but that is the overwhelming focus. Philosopher Peter Singer started the original idea in 2013 and bought into it big time.^[4] Effective Altruism is also pushed by Bay Area technolibertarians, and artificial intelligence existential risk groups, including MIRI. The latter, of course, consider themselves an obvious beneficiary — if not the obvious beneficiary (i.e., self-dealing). After EA began gaining traction, Singer became very wary about EA, stating in 2021:^[5]

The sales pitch is that, if you're going to try make the world a better place for other people, you should try to do the best possible job you can. If you had the choice between helping a local community theater group put on a show or saving African children from malaria, the right thing to do is, of course, to save the children. People face dilemmas like this in real life whenever they donate money to charity: if you're not donating to the most cost-effective charities that you can, you fail at utilitarianism. (It's impossible not to fail at utilitarianism, but you can fail less hard.)

It is important to remember that EA invented neither the concept of charity, nor the concept of evaluating charities — though some EAs behave as though they did. Beware of EAs equivocating by responding to criticisms of the EA subculture's behaviours with advocacy of the value of charity or evaluating charities in general.

How EAs evaluate charities[edit]

In the ideal case, EAs leave actually evaluating how good charities are to dedicated organisations set up for that purpose — charity evaluators, such as GiveWell and Giving What We Can (GWWC). GiveWell and GWWC tend to rate charities in a quasi-utilitarian way, using a combination of the best available published evidence for the interventions, and asking lots of questions of the charities they rate relating to things like checking that the interventions actually work (auditing), room for more funding, and whether adding more funding would do the same amount of good, more good, or less good. Overhead is also considered: however, overhead is not regarded as a terrible thing if it invests the effectiveness of the work (monitoring programs being a notable example). They are preferred by EAs over existing charity evaluators such as Charity Navigator — Charity Navigator just looks at the percentage a charity spends on administrative and fundraising overheads and pays no attention to whether what the charity is doing is effective, or how effective it is.

Of course, then there's donations to MIRI, but MIRI appear to be special-cased for subcultural reasons.

GiveWell[edit]

However, GiveWell has partnered with billionaire Facebook co-founder Dustin Moscovitz and his wife's charitable foundation in a joint initiative called the Open Philanthropy Project, and in this initiative they have been accused of casting aside analytical rigour, and have been accused of bias.^[6]

GiveWell has also recommended that people spam the Against Malaria Foundation (AMF) with all^{[note 1]} the money they have set aside to donate, on the grounds that they think it's the best charity, even at the risk of exhausting the AMF's room for more funding, amongst other dubious decisions.^{[citation needed]}

Effective altruists have criticized GiveWell for being too strict in their criteria which leaves UNICEF out of their list of recommended charities because UNICEF focuses on too many interventions which makes it harder for GiveWell to evaluate their effectiveness.^[7] This is despite the fact that UNICEF engages in many cost-effective interventions, such as providing vaccines.

Origins[edit]

Singer at an EA conference in 2015

Yudkowsky (left) at the Singularity Summit in 2007

The philosophical underpinnings mostly come from philosopher Peter Singer, particularly his 1972 essay Famine, Affluence, and Morality. He argues in this essay that affluent people are morally obligated to donate far more of their income to humanitarian causes than is considered normal in Western culture. This did not start the effective altruism subculture, but once it was going, he joined in enthusiastically.

The effective altruism subculture — as opposed to the concept of altruism that is effective — originated around LessWrong.^[8] The earliest known use of the term was in the form "effectively altruistic" by user "Anand" in a 2003 edit on the wiki of the singularitarian Shock Level 4 mailing list, a predecessor of LessWrong run by Eliezer Yudkowsky.^[9] Anand's article argued that donating to the Singularity Institute (now known as MIRI) is more effective than donating to prevent the spread of HIV/AIDS, even though the latter may be more emotionally compelling. Later, the term was used in the form "effective altruist" by Yudkowsky himself in his 2007 blog post Scope Insensitivity, arguing against sentimentality and for utilitarian calculation in charity:^[10]

If you want to be an effective altruist, you have to think it through with the part of your brain that processes those unexciting inky zeroes on paper, not just the part that gets real worked up about that poor struggling oil-soaked bird.

Other names were used, e.g. "efficient charity" in 2010,^[11] but the movement eventually settled on the name "effective altruism" by 2012.^[12][8]

Earning to give[edit]

People that call themselves effective altruists commonly endorse the "earning to give" approach, at least for those who have, or might be able to get, well-paid jobs. At its most hardcore, "earning to give" means getting the highest-paying job that one can and then donating as much of it as possible (up to some threshold, for sanity's sake). After all, you can get more done by paying a bunch of other people to solve problems for you than you can do all on your own, right?^[13] One story about EA originally did not give much heed to the morality or potential for harm of the job itself, recommending "trading in quantitative hedge funds" in 2014-2015,^[14] later adding a caveat in 2017:^[13]

But still retaining "quantitative trading" as a career option.^[13] This does not appear to be a harmless career choice at all in light of the quantitative trading that helped lead to the Great Recession in 2009, or quantitative trading that led to the bankruptcy of the FTX cryptocurrency exchange in 2022.^[15] Disregarding the potential harm that one's job causes for the purposes of philanthropy, makes the philanthropy either a smokescreen, or the whole process one of consequentialism (ends justifying the means).

In practice, people will not always take (or keep) the highest-paying job they can, for a variety of reasons including commute time, company culture, working hours, the employer's attitude to diversity, work-related stress, and whether the management are perceived to treat employees well or badly. However, 80,000 Hours, an organisation dedicated to giving career advice to wannabe effective altruists, published a blog post claiming that research showed that, depending on the type of stress, stress at work wasn't necessarily a big deal anyway and in some cases, people should consider just sucking it up and maybe "reframe stress as opportunity", in the interests of saving more children from malaria.^[16]

Also, in practice nobody literally donates "as much as possible," an unrealistic standard which would presumably mean forgoing any kind of luxuries and a curtailed social life, at least after securing a long-term relationship — and which would still leave the awkward question of whether one's kids should be brought up in near-poverty. (The powerful human instinct towards protection of one's own offspring would tend to mitigate against such thinking when it came down to it — and if not, there's always social services.) One EA organisation, Giving What We Can, promotes a suggested amount of 10% of one's entire working lifetime income, spread over a working lifetime. Although this is easily achievable by generously-compensated San Francisco Bay Area software engineers, and (as even Giving What We Can recognises) not achievable by students struggling to get by on student loans, some in the movement seem curiously blind to the fact that not everyone who has a job might be able to part with 10% of their entire income (what amounts to a voluntary flat tax — punitive for the poor and easy for the rich). Some — not all of them millionaires — even pledge to give much more than 10% of their income. It is unclear whether this behaviour is, on balance, inspirational, or whether it acts to drive away potential donors, activists and charity workers who might feel that this is a movement of exclusively privileged people that is remote from their lives and concerns.

Compounding the problem, effective altruism is regularly conflated, even inside the movement, with:

• Giving What We Can, even though not all people who identify as "effective altruists" have pledged to donate 10% of their income or are planning to do so
• Utilitarianism, even though not all effective altruists are utilitarians
• Supporting everything that everyone in the movement does, even though that would be arguably self-contradictory (see below)

EA organisations regularly conduct research into what brings people into the EA movement, but no formal research seems to have been done into what drives some people away from EA. The thinking of many EAs is that effective altruism is so obviously right, only people who were somehow in fundamental disagreement with EA values like doing nice things, and doing more and better things rather than fewer and worse things, would even consider not joining the movement...

Mosquito nets versus artificial intelligence risk[edit]

Karnofsky visiting the NGO Seva Mandir in Rajasthan, India in 2010

The ideas have been around a while, but the current subculture that calls itself Effective Altruism got a big push from Machine Intelligence Research Institute (MIRI) and its friends in the LessWrong community, many of whom considered MIRI obviously the most effective charity in the world.^{[note 2]} However, unfortunately for MIRI, EA charity guide GiveWell subsequently rated donations to them as actually worse for their project (addressing the threat to humanity posed by hypothetical future advanced Artificial Intelligence technology) than not donating, with GiveWell's Holden Karnofsky stating in 2012, "I do not believe that these objections constitute a sharp/tight case for the idea that SI's work has low/negative value; I believe, instead, that SI's own arguments are too vague for such a rebuttal to be possible." GiveWell, unlike LessWrong and MIRI, primarily promotes charities focused on improving health in the developing world. GiveWell's criticism of MIRI argued that MIRI's focus on supposedly trying to save the world and create "Friendly AI" amounted to a form of Pascal's Mugging — promising enormous benefits, even though the probability of actually receiving those benefits is tiny.^[19][20]

This is not the only example. Reducing animal suffering is an important cause for some of the movement,^[21] but some people have unusual ideas on how to do this. One prominent effective altruist has put up for discussion (since retracted) on his blog the idea of destroying nature in order to reduce wild animal suffering.^[22] Was it satire or serious? One may never know. In fact, some members of the Effective Altruism movement identify as "negative utilitarians", meaning that preventing suffering is the only thing that matters. However, this philosophy seems to imply that we should be willing to destroy the entire world to prevent one person from suffering a pinprick,^[23] or at least anti-natalism.

These examples represent internal tensions around a key concept of EA, longtermism, that was developed by William MacAskill and Toby Ord.^[24] MacAskill defines longtermism as:

…the idea that positively influencing the longterm future is a key moral priority of our time. Longtermism is about taking seriously just how big the future could be and how high the stakes are in shaping it. If humanity survives to even a fraction of its potential life span, then, strange as it may seem, we are the ancients: we live at the very beginning of history, in the most distant past. What we do now will affect untold numbers of future people We need to act wisely.^[25]:4-5

This sounds wise and benevolent, but if predicting the future is often highly fraught, even a few years out. Without paying attention to immediate world problems, one can easily waste money and perhaps create unintended negative consequences.

Despite the many and varied differences of opinion within the EA movement, those that remain in the movement tend not to spend too much time arguing about fundamental "cause selection" issues (whether to donate to AI risk, global health, poverty or animal causes) — and even when they do, such discussions tend to remain relatively civil and non-rancorous. Part of the reason for this is that all EAs are in favour of "growing the pie" of EA supporters at this point in time, and most of them recognise that rancorous discussions would impede that goal. Although ideas about targeting growth differently have been mooted, such as focusing more on trying to recruit the rich (by hard-headed pragmatists) or women and ethnic minorities (by social justice people) or people who don't speak English (by people who think outside the English-speaking world), no-one is so pessimistic about their favoured EA cause area that they think that growing the pie won't gain their preferred cause area more EA recruits.

However, one EA has argued that this polite truce doesn't make sense, because if people think their cause is vastly better, they should be spending a lot of their time trying to persuade people of that.^[26] Scott Alexander, an EA supporter,^[27] has counter-argued, based on his extensive personal (and often unsuccessful) experience of arguing with people who are sceptical about AI risk as a cause, that repeated arguments of this kind at EA meetups would be tiring, repetitive, and unpleasant.^[26] This is not to say that Alexander does not advocate for AI risk reduction — however, he prefers to write long blog posts where he can assemble his arguments and evidence and engage in an extensive, uninterrupted written monologue.^[28]

Where "Effective Altruists" actually send their money[edit]

Wytham Abbey: effective self-dealing

According to William MacAskill of "The Effective Altruism Blog", effective altruists currently tend to think that the most important causes to focus on are global poverty, factory farming, and the long-term future of life on Earth.^[29] In practice, this amounts to complaining when people try to solve local problems, feeling bad when people eat hamburgers,^{[note 3]} and sending money to Eliezer Yudkowsky, respectively.

The effective-altruism.com 2014 Survey of Effective Altruists was self-selected and non-random, but includes a list of how many respondents said they donated to various organisations:^[30][31]

The organisations listed in boldface above are directly linked to EA, i.e. founded by, staffed by, and/or would not exist without EA enthusiasts. When one is essentially self-dealing like this, it's hard to argue that it's actual altruism (i.e., "disinterested and selfless concern for the well-being of others"^[35]).

Cultishness[edit]

EA has been accused of being a cult, and it does have a few of the hallmarks of a cult:^[36][37]

• Obsession with hypothetical existential risk such as Roko's basilisk and longtermism^[37]
• Groupthink centered around leaders Sam Bankman-Fried's FTX and Sam Altman's OpenAI and Nick Bostrom's Future of Humanity Institute in particular^[37]
• Intentional use of equations in rhetoric to baffle followers and detractors alike^{[37][38][39][40]}
• The flawed premise of utilitarianism that underlies EA
• Bostrom's apparently viewing himself as a superhero^[37][41]

The Economist quoted one unnamed effective altruist as saying, "Don’t ever say ‘People sometimes think EA is a cult, but it’s not."^[37] The article next said that EA may not be a cult but rather "kind of church – one that has become increasingly centralised and controlled over time."^[37] An edited book on EA and religion has even been written.^[42]

Risks[edit]

Robin Hanson points out that people accumulate knowledge and wisdom as they get older and may change their minds about important things as a result. He therefore advises effective altruists — those of them who are young, anyway, which is most of them — to do nothing now and save money for later, because they might change their mind about where to give that money. Hanson, who specialises in the study of allegedly hypocritical human behaviour ("X is not really about Y!") argues that effective altruists are prone to irrationally give now rather than later, to signal sincerity to their fellow effective altruists. When asked whether it is not better to give now while we still can, because our future selves might spend the money on e.g. putting our children through university, he responded "Maybe that's the right thing to do! Why do you distrust your future self so much?"^[43]

Like activism and do-gooding generally, for high-scrupulosity people, going overboard with EA can be dangerous. It can lead to burnout from overwork and/or neglecting one's own needs and/or those of one's family. It's worth remembering that, "effective" as it may be to buy a bed net for a child in Africa, people close to you also have needs of various sorts, which can often most "effectively" be met by you.

A significant number of EAs advocate giving large portions (10%+) of one's income away on a continuous basis, but it is important to remember that one's circumstances may change — for example, one may lose one's job or encounter a health crisis — so it is worthwhile considering saving some money in case you need it. One can always give away that saved money later, or change one's mind if one decides that the money is really need for oneself.

Excessive moralising about EA can also cause one to — like a kind of inverse Dale Carnegie — lose friends and fail to influence people. Arguably, persuading other people to give to good causes is best approached in an upbeat "look what you could achieve" way, rather than trying to guilt-trip people. (The latter is probably more likely to work on people who were already high-scrupulosity and thus more susceptible to EA ideas in the first place — so the value of "converting" such people to EA by guilt-tripping them could well be less than you might think, because they might have ended up being converted anyway.)

In the (unlikely) worst case scenario, you could lose all your non-EA friends through being seen as extremely preachy and arrogant, then later become financially ruined through a chance accident or illness leaving you unable to work, have no savings to fall back on — and then receive no help whatsoever from your EA friends despite all the past good you have done, because helping you is not an "effective" cause. This scenario is probably unlikely to pan out this way in practice though. Probably.

Exemplars[edit]

Sam Bankman-Fried[edit]

Bankman-Fried remotely attending the MIT Bitcoin Expo 2022

Sam Bankman-Fried (SBF) (1992–), owner of cryptocurrency exchange FTX until bankruptcy in November 2022, was a vocal (with a voice that could drive a monk to insanity, no less) proponent of EA and had pledged to give away his fortune to charity, signing the Giving Pledge.^{[46][47][note 8]} It's unclear whether SBF ever gave money to charity, though he did make a lot of political contributions (likely to fend off crypto regulation)^[47] allegedly illegally using other people's money.^[48] SBF's crypto empire spectacularly came crashing down in November 2022, due to his self-dealing and high-risk (leveraged) gambles and he did claim to have given a lot of moola to Jane Street Capital, both founded by his mentor (so basically self-dealing).^[49] After the crash, which left SBF with essentially nothing, it was revealed that he was living a lavish and debauched lifestyle in the Bahamas^[50] while he was hypothetically donating to charity at some point in the distant future. Trial testimony and his own statements indicated that SBF had an extreme desire for risk, even to the point of risking other people's lives without their consent,^[51][52] something that would seem to be antithetical to the idea of altruism. The bookkeeping for FTX, as it turned out, was close to non-existent^[44] (Was it a six-paragraph blog post?).

In November 2023, SBF was convicted of 2 counts of wire fraud, 4 counts of conspiracy to commit fraud and 1 count of conspiracy to commit money laundering in connection with FTX and the affiliated Alameda Research cryptocurrency trading fund.^[54] In 2024, SBF was sentenced to 25 years in prison.^[55]

Lest one think that SBF was somehow an aberration within EA, both of SBF's parents were involved with FTX. His parents are both law professors at Stanford University, with his father Joseph Bankman specializing in tax law, and his mother Barbara Fried specializing in legal ethics and wrote a book that supports consequentialist philosophies!^[56] SBF grew up immersed in utilitarianist thought, and fully embraced utilitarianism in a blog post.^[57]:81-82

Both of SBF's parents have been sued by FTX creditors to recover funds, but it is unclear if they will also face criminal charges.^[58]

Ben Delo[edit]

Ben Delo (1984–) was also courted by MacAskill to join the EA movement.^[59][60] Delo became a member of MacAskill's Giving What We Can,^[61] and signed the Giving Pledge.^[46] Subsequently Delo was convicted of willfully failing to implement anti-money laundering measures at the crypto exchange BitMEX, sentenced to 30-months probation and ordered to pay a $10 million fine.^[62]

The poor and cryptocurrency[edit]

Given that EA has at times been about the rich helping the poor, it is noteworthy to look at the cryptocurrency entanglements of EA and the effect of EA on the poor. Besides the crypto debacles of SBF and Delo, Elon Musk has promoted cryptocurrency^[63][64] and EA,^[65] and MacAskill advised SBF before he created FTX. MacAskill also promoted the idea of intentionally working for an immoral organization so as to earn money to donate to charity. MacAskill gave an example to justify this, that of Nazi Oskar Schindler who saved Jews from the Holocaust.^[66] This was an example of a Nazi who later happened to rescue Jews, not of someone who became a Nazi so that he could rescue Jews. What MacAskill is trying to justify is the latter, e.g. help destroy the planet by working in the fossil fuel industry while promising to donate some money to an anti-climate change charity.

Cryptocurrency has been directly and indirectly associated with various forms of impoverishment through criminality:

• Pump-and-dump schemes^[67]
• Theft of poor people's crypto from the Axie Infinity game^{[68][57]:121-128}
• Human trafficking and enslavement of poor people, then forcing them to con others into fraudulent crypto investments that depend on Tether (text flirting followed by enticement into fraudulent crypto investments, a method known as "pig butchering").^[57]:172-199

Gallery[edit]

• 

  Delo, promoting Bitcoin in 2018

• 

  MacAskill in 2018

• 

  Ord in 2019

See also[edit]

For those of you in the mood, RationalWiki has a fun article about Effective altruism.

• ChatGPT — several members of OpenAI's governing board have had ties to EA.^[69]
• Pascal's wager

External links[edit]

• Longtermism Hub — critiques of longtermism

Notes[edit]

References[edit]