View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by UCL Discovery

Stop the Robot Apocalypse Amia Srinivasan

Doing Good Better: Effective and a Radical New Way to Make a Difference by William MacAskill. Guardian Faber, 325 pp., £14.99, August, 978 1 78335 049 0

hilosophy, Wittgenstein said, lecturer at Oxford. As graduate students ‘leaves everything as it is’. It sounds MacAskill and his friend commit­ Plike a complaint, but actually it was a ted themselves to donate most of their fut­ recommendation. Philosophy at its best, ure earnings to (in MacAskill’s case Wittgenstein thought, resists the scientific anything above £20,000, in Ord’s £18,000), impulse to treat the world as a theoretical and set themselves the task of figuring out construct. It is not a view shared in the how to make best use of the money they main by contemporary philosophers. What had pledged. The result was Giving What is philosophy supposed to do if not theor­ We Can, a charity that encourages people ise? At the same time most philosophers to hand over at least 10 per cent of their are happy to leave everything as it is in a ­future incomes for philanthropic purposes, more prosaic sense: that is, by not really and advises them on how to get the most changing anything. Philosophers may talk out of their money. Since the charity was about justice or rights, but they don’t often founded in 2009 it has received more than try to reshape the world according to their $400 million in pledges, much of it from ideals. Maybe that’s for the best. Philo­ young philosophers. In 2011, MacAskill set sophers have a tendency to slip from sense up 80,000 Hours (the name refers to the into seeming absurdity: a defence of abort­ number of hours the average person works ion ends up defending infanticide; an argu­ over a lifetime), a charity that helps people ment for vegetarianism turns into a call make career choices with the aim of maxi­ ­for the extermination of wild carnivores. mising social benefit; it raised eyebrows A new generation of moral philosophers early on by advising graduates to become is determined to break with this tradition philanthropic bankers rather than NGO of ineffectuality. The goal of the ‘effective workers. The two organisations are incorp­ altruists’ is not only to theorise the world, orated as the Centre for , but to use their theories to leave the world based in Oxford, and are in the van of a a better place than they found it. Their global movement, encompassing groups leader is William MacAskill, a 28-year-old such as GiveWell (founded by two hedge-

1 16 September 2015 fund managers at around the same time as who don’t need dialysis. Maybe this is be­ MacAskill and Ord started their work), The cause dialysis isn’t as bad as we think. Or Life You Can Save (founded by the philoso­ maybe it’s because dialysis is so awful that pher ), (found­ you forget just how much better your life ed by the cofounder Dustin Mos­k­ was without it.) To calculate whether, given ovitz and his wife, Cari Tuna, who have the choice, it would be better to cure the pledged to give away most of their money), blind person or improve the life of the Aids Animal Charity Evaluators (an 80,000 Hours patient, you must take into account the in­ spin-off ) and the Open Philanthropy Pro­ crease in both life quality and life expectancy ject (a collaboration between GiveWell and that would be caused by the intervention. Good Ventures). Giving a 40-year-old Aids patient antiretro­ In , MacAskill sets out virals would give her a 40 per cent jump in the thinking behind effective altruism. His life quality (from 50 per cent to 90 per cent) main claim, familiar from the utilitarian trad­ for five years, and also would give her an ad­ ition out of which the movement emerges, ditional five years of life at 90 per cent, for is that we should seek not only to do good, a total of 6.5 Qalys ((0.4 x 5) + (0.9 x 5) = but to do the most good we can. To do that 6.5). Curing a blind 20-year-old, assuming we need empirical research – research his he lives to be 70, would increase his quality organisations provide – into the amount of of life from 40 per cent to 100 per cent for good created by various different charities, 50 years, for a total of 30 Qalys. So curing types of consumption, careers and so on. the blind person is more valuable, in terms MacAskill proposes that ‘good’, here, can be of Qalys, than giving drugs to the Aids pat­ understood roughly in terms of quality-­ ient. Thinking in terms of Qalys makes it adjusted life-years (Qalys), a unit that allows possible to compare that which seemingly welfare economists to compare benefits of cannot be compared: blindness with Aids; very different sorts. One Qaly is a single increases in life expectancy with increases year of life lived at 100 per cent health. Ac­ in life quality. Qalys free us from the spec­ cording to a standardised scale, a year as an ificity of people’s lives, giving us a universal Aids patient not on antiretrovirals is worth currency for misery. 0.5 Qalys; a year with Aids lived on anti­ However, when deciding what to do – retrovirals is worth 0.9 Qalys. A year of life what job to get, which charity to donate to, for a blind person is worth 0.4 Qalys; a year whether to buy Fairtrade or not – it isn’t of life as a non-blind, otherwise healthy enough, according to MacAskill, to think in person is worth 1 Qaly. (These numbers are terms of Qalys. We must also think both based on self-reporting by Aids patients and marginally and counterfactually. The idea blind people, which raises some obvious that value should be measured on the worries. For example, dialysis patients rate ­margin is familiar from economics; it’s what their lives at 0.56 Qalys – significantly high­ explains the fact that, say, heating repair­ er than the 0.39 Qalys predicted by people men make more money than childcare work­

2 16 September 2015 ers. Presumably childcare workers produce ­doctor in the developing world you’d save more total value than heating repairmen, the equivalent of 300 lives (or 10,950 Qalys, but because the supply of childcare work­ at an average 36.5 Qalys per life) over a 40- ers is greater than the supply of good repair­ year career. Yet if you didn’t take that job, men, we will pay more for an additional re­ someone else probably would; they may pairman than an additional childcare work­ not save quite as many lives as you, but they er. The average value of a childcare worker would save most of them. Meanwhile you might be higher than the average value of could quit medicine, take a high-paying a repairman, but the repairman has the ­finance job and donate most of your salary greater marginal value. (Another way of put­ each year to the most effective charities. ting this: coffee might be really important MacAskill estimates that you can expect to to you, but if you’ve already had three cups save a life with $3400 by donating to the you’re probably not going to care as much Against Foundation, which pro­ about a fourth.) Similarly, MacAskill reas­ vides insecticide-coated bed nets to poor ons, when giving money to charity or de­ families. A financier who worked for 40 years, ciding, say, whether to become a doctor, donating $50,000 a year to the Against we should focus on marginal rather than Malaria Foundation, could expect to save average value. The average doctor in the de­ around 580 lives – lives, significantly, that veloped world helps save a lot of lives, but would not have been saved otherwise. Qaly the marginal doctor – because the supply of thinking frees us from considering the doctors is large, and most of the life-saving specificity of whom we are helping; marg­ work is already covered – doesn’t. The marg­ inal and counterfactual thinking frees us inal doctor in the developing world has from the specificity of ourselves. What greater value, since the supply of doctors is matters isn’t who does the good, only that lower there. MacAskill estimates that a good is done. doctor practising in a very poor country But don’t many lucrative careers have adds about a hundred times as much marg­ bad social effects? Up until recently Mac­ inal value (measured in Qalys) as a doctor Askill argued that such effects were moral­ practising in the UK. (In general, MacAskill ly irrelevant, again by counterfactual reas­ says, a pound spent in a poor country can oning: if you didn’t take the banking job do one hundred times more good than it someone else would, so the harm would be can in a rich one, a heuristic he calls the done anyway. (In an academic paper pub­ ‘100x Multiplier’.) lished last year, he compares a philanthrop­ But before signing up for a medical ic banker to Oskar Schindler, who provided ­career in sub-Saharan Africa you should be munitions to the Nazis as a means of saving careful, MacAskill warns, to evaluate the the lives of 1200 Jews; if Schindler hadn’t counterfactual. That is, you have to ask your­ manufactured the arms, some other Nazi self what would happen if you didn’t be­ would have, without saving any Jewish lives.) come a doctor at all. Let’s say that as a More recently MacAskill and his team at

3 16 September 2015 80,000 Hours have backed away from this mostly to middlemen rather than farmers, ‘replaceability thesis’, conceding that it’s and when it doesn’t, it benefits farmers in harder than they initially thought to eval­uate relatively rich countries: because Fairtrade the counterfactuals. For example, there’s standards are hard to meet, most Fairtrade good economic reason to think that going coffee production comes from Mexico and into banking really does increase the total Costa Rica rather than, say, Ethiopia, where number of bankers, and doesn’t simply the marginal pound would go much fur­ change who does the banking. ­MacAskill ther. The green value of buying locally says he no longer recommends that people grown food is overblown, too, since trans­ go into banking, or at least not the parts of port accounts for only 10 per cent of the it that he thinks cause direct harm: creating carbon footprint of food, while 80 per cent risks that will be borne by unsuspecting of it is generated in production; tomatoes taxpayers, or selling pro­ducts that no prop­ grown in the UK can have five times the car­ erly informed person would buy. Instead bon footprint of tomatoes shipped from 80,000 Hours now encourages people to Spain because of the energy required to take what it sees as morally­ neutral or posit­ hothouse them. If you’re really committed ive jobs: quantitative hedge-fund trading, to minimising your carbon footprint, Mac­ management consulting, technology start- Askill recommends donating to the carbon ups. (You can take a careers quiz on the offsetting charity Cool Earth; he estimates 80,000 Hours website; I was told to become that the average American could offset all a consultant, because of its earning-to-give his carbon emissions by donating $105 a potential and the general­ business educ­ year. There isn’t much point in unplugging ation it provides. When I changed my ans­ your electricals, either: leaving your mobile wers to say that I was bad at maths I was phone charger plugged in for a whole year told to go into politics.) contributes less to your carbon footprint The results of all this number-crunching than one hot bath. are sometimes satisfyingly counterintuit­ ive. Deworming has better educational out­ oing Good Better is a feel- comes among Kenyan schoolchildren than good guide to getting good done. It increasing the numbers of textbooks or Ddoesn’t dwell much on the horrors teachers. If you want to improve animal of global inequality, and sidesteps any dia­ welfare, it’s better to stop eating eggs than gnosis of its causes. The word ‘oppression’ beef, since caged layer hens live worse lives appears just once. This is surely by design, than farmed cows, and because eating eggs at least in part. According to MacAskill’s­ consumes more animals than eating beef: moral worldview, it is the consequences of the average American consumes 0.8 layer one’s actions that really matter, and that’s hens but only 0.1 beef cows per year. Buy­ as true of writing a book as it is of donating ing Fairtrade goods can be worse than buy­ to charity. His patter is calculated for max­ ing regular goods, since the extra cost goes imal effect: if the book weren’t so cheery,

4 16 September 2015 MacAskill couldn’t expect to inspire as altruism is just the latest instance. much do-gooding, and by his own lights Yet there is no principled reason why ef­ that would be a moral failure. (I’m not say­ fective altruists should endorse the world­ ing it doesn’t work. Halfway through read­ view of the benevolent capitalist. Since ef­ ing the book I set up a regular donation to fective altruism is committed to whatever GiveDirectly, one of the charities Mac­Askill would maximise the social good, it might for endorses for its proven efficacy. It gives ­un example turn out to support anti-capitalist conditional direct cash transfers to poor revolution. And although MacAskill focuses households in Uganda and Kenya.) on health as a proxy for goodness, there is But the book’s snappy style isn’t just a no principled reason, as he points out, why strategic choice. MacAskill is evidently com­ effective altruism couldn’t also plug values fortable with ways of talking that are fam­ like justice, dignity or self-determination iliar from the exponents of global capital­ into its algorithms. (There’s also no reason ism: the will to quantify, the essential why one couldn’t ‘earn to give’ to help ­comparability of all goods and all evils, the ­radical causes; Engels worked at a mill in obsession with productivity and efficiency, Manchester to support Marx’s writing of the conviction that there is a happy converg­ Capital.) Effective altruism has so far been a ence between self-interest and morality, rather homogenous movement of middle- the seeming confidence that there is no class white men fighting through ­crisis whose solution is beyond the ingen­ largely conventional means, but it is at uity of man. He repeatedly talks about phil­ least in theory a broad church. Indeed one anthropy as a deal too good to pass up: ‘It’s element of the movement is turning its at­ like a 99 per cent off sale, or buy one, get 99 tention towards what members like to call free. It might be the most amazing deal ‘systemic change’, taking up political ad­ you’ll see in your life.’ There is a seemingly vocacy on issues ranging from factory farm­ unanswerable logic, at once natural and ing to reform. Even in these magical, simple and totalising, to both cases, the numbers are what matter. Mac­ global capitalism and effective altruism. Askill describes how he helped an Oxford That he speaks in the proprietary language PPE student work out whether or not she of the illness – global inequality – whose should get into electoral politics. He calcul­ symptoms he proposes to mop up is an ates that historically, the odds of a politic­ ­irony on which he doesn’t comment. Per­ ally ambitious Oxford PPE student becom­ haps he senses that his potential followers ing an MP have been one in thirty (he notes – privileged, ambitious millennials – don’t that this reflects ‘some disappointing facts want to hear about the iniquities of the about political mobility and equal repre­ ­system that has shaped their worldview. sentation in the UK’). Applying some con­ Or perhaps he thinks there’s no irony here servative estimates of the resources an aver­ at all: capitalism, as always, produces the age MP gets to control, he prices the marg­ means of its own correction, and effective inal expected value of the student’s run­

5 16 September 2015 ning for Parliament at £8 million, which like everyone else, come up against the fact turns out to be high enough, compared with that the world is messy, and like everyone the expected value of other careers she else who wants to make it better they must might pursue, to justify the move into pol­ do what strikes them as best, without any itics. It’s not clear that anyone with less final sense of what that might be or any conventional political ambitions would get guarantee that they’re getting it right. the same pass. What’s the expected marg­ More worrying than the model’s inabil­ inal value of becoming an anti-capitalist ity to tell us anything very useful once we revolutionary? To answer that you’d need move outside the circumscribed realm of to put a value and probability measure on controlled intervention is its susceptibility achieving­ an unrecognisably different world to being used to tell us exactly what we – even, perhaps, on our becoming unrecog­ want to hear. A three-day conference, ‘Ef­ nisably different sorts of people. It’s hard fective Altruism Global’, was held this sum­ enough to quantify the value of a phil­ mer at Google’s headquarters in Mountain anthropic intervention: how would we go View, California. While some of the ses­ about quantifying the consequences of rad­ sions focused on the issues closest to Mac­ ically reorganising society? Askill’s heart – cost-effective philanthropy, MacAskill seems to think there is no moral global poverty, career choice – much of it calculation that can’t be made to fit on the was dominated, according to Dylan Mat­ back of his envelope; any uncertainty we thews, who was there and wrote about it for might have about precise values or prob­ Vox, by talk of existential risks (or x-risks, as abilities can be priced into the model. (His the community calls them). An x-risk, as doctoral dissertation was on the model­ defined by the Oxford philosopher Nick Bos­ ling of moral uncertainty.) But the more trom, who popularised the concept, is an uncertain the figures, the less useful the event that would ‘permanently and drastic­ calculation, and the more we end up rely­ ally curtail humanity’s potential’ – total ing on a commonsense understanding of ­annihilation is the obvious case. Given the what’s worth doing. Do we really need a number of people who might live in the ­sophisticated model to tell us that we ­future if not for such an event – Bostrom shouldn’t deal in subprime mortgages, or ­estimates the figure at 1052, assuming that that the American prison system needs fix­ we master interstellar travel and the up­ ing, or that it might be worthwhile going loading of human minds to computers – into electoral politics if you can be confid­ the expected value of preventing an x-risk ent you aren’t doing it solely out of self- dwarfs the value of, say, curing cancer or interest? The more complex the problem preventing genocide. This is so even if the effective altruism tries to address – that is, prob­ability of being able to do anything the more deeply it engages with the world about an x-risk is vanishingly small. Even if as a political entity – the less distinctive its Bostrom’s­ 1052 estimate has only a 1 per contribution becomes. Effective altruists, cent chance of being correct, the expected

6 16 September 2015 value of reducing an x-risk by one billionth Askill does not address the deep sources of one billionth of a percentage point (that’s of global misery – international trade and 0.0000000000000000001 per cent) is still a finance, debt, nationalism, imperialism, hundred billion times greater than the value­ racial and gender-based subordination, war, of saving the lives of a billion people liv­ environmental degradation, corruption, ing now. So it turns out to be better to try exploitation of labour – or the forces that ­to prevent some hypothetical x-risk, even ensure its reproduction. Effective altruism with an extremely remote chance of being doesn’t try to understand how power works, able to do so, than to help actual living except to better align itself with it. In this people.­ sense it leaves everything just as it is. This X-risks could take many forms – a met­ is no doubt comforting to those who enjoy eor crash, catastrophic global warming, the status quo – and may in part account for plague – but the one that effective altruists the movement’s success. like to worry about most is the ‘intelligence Yet behind MacAskill’s cheery exhort­ explosion’: taking over ation to invest in anti-malarial nets lies the world and destroying humanity. Their a moral philosophy that really is radical. favoured solution is to invest more money In 1972 Peter Singer published his paper in AI research. Thus the humanitarian­ ‘Famine, Affluence and Morality’, a classic logic of effective altruism leads to the of contemporary , in which he ­conclusion that more money needs to be compares a Westerner who spends money spent on computers: why invest in anti- on luxuries rather than don­ating it to the malarial nets when there’s a robot apoc­ developing world to some­one who walks by alypse to halt? It’s no surprise that effective a drowning child rather than get his clothes altruism is popular in Silicon Valley: PayPal muddy. We can all agree that the second founder Peter Thiel, developer Jaan case is morally abhorrent, but not everyone Tallinn and Tesla CEO Elon Musk are all has the same qualms about the first. What’s major financial supporters of -x risk re­ the difference? Does it really matter, Singer search.* Who doesn’t want to believe that asks, that a child in the developing world is their work is of overwhelming humanitar­ thousands of miles away rather than in ian significance? front of us? If not, then all of us who don’t The subtitle of Doing Good Better promises merely subsist, who spend money on our­ ‘a radical new way to make a difference’; selves when it would be worth significantly one of the organisers of the Googleplex more to someone else, are morally im­ conference declared that ‘effective altruism plicated in murder, just like the man who could be the last social movement we ever allows the child to drown. The vast scale of need.’ But effective altruism, so far at least, global inequality – if your income is more has been a conservative movement, calling than £34,000 per year, adjusted for pur­ us back to where we already are: the world chasing power, you’re in the global 1 per as it is, our institutions as they are. Mac­ cent – means that even the smallest lux­

7 16 September 2015 uries (going to the cinema, a second pair of less how it could be the last social move­ shoes, a drink at the pub) may be morally ment we’ll ever need. unacceptable. Effective altruism takes up the spirit of more pressing objection to utilit­ Singer’s argument but shields us from the arianism is not that it demands too full blast of its conclusion; moral indict­ Amuch, but that it demands the ment is transformed into an empowering wrong things, the things that constitute us investment opportunity. Instead of down­ as humans: our personal attachments, loy­ grading our lives to subsistence levels, we alties and ident­ifications. On the utilitarian are encouraged to start with the traditional view, a pound spent without maximal effect tithe of 10 per cent, then do a bit more each is a pound spent immorally. Luxuries are year. Thus effective altruism dodges one of naturally rul­ed out, but so is spending on the standard objections to utilitarianism: worthwhile causes to which you might feel that it asks too much of us. But it isn’t clear some person­al affinity. Here MacAskill how the dodge is supposed to work. Mac­ agrees: to choose to donate to a relatively Askill tells us that effective altruists – like cost-ineffective charity just because it’s utilitarians – are committed to doing the close to your heart – the local soup kitchen, most good possible, but he also tells us that or a seeing-eye dog charity in honour of a it’s OK to enjoy a ‘cushy lifestyle’, so long blind relative (it costs £32,400 to train one as you’re donating a lot to charity. Either seeing-eye dog and its owner) – is wrong. ­effective altruism, like utilitarianism, de­ How far should the effective altruist go mands that we do the most good possible, with this logic? If you’re faced with the or it asks merely that we try to make things choice between spend­ing a few hours con­ better. The first thought is genuinely radic­ soling a bereaved friend, or earning some al, requiring us to overhaul our daily lives money to donate to an effective charity, the in ways unimaginable to most. (Singer re­ utilitar­ian calcul­us will tell you to do the peats his call for precisely such an overhaul latter. If effective altruists really are com­ in his recent book , mitted to doing the most good, they should and Larissa MacFarquhar’s Strangers Drown- say the same. If however they are merely ing is a set of portraits of ‘extreme altruists’ committed to doing a lot of good, then they who have answered the call.†) The second will say that you can stay with your friend so thought – that we try to make things better long as you’re doing sufficient good else­ – is shared by every plausible moral system where. But even this more moderate view and every decent person. If effective altru­ misconceives the situation. You should stay ism is simply in the business of getting us and console your friend not because you’ve to be more effective when we try to help already met your do-gooding quota, but be­ others, then it’s hard to object to it. But in cause it’s your friend that is in distress. This that case it’s also hard to see what it’s offer­ is also the reason you shouldn’t deal in sub­ ing in the way of fresh moral insight, still prime mortgages or make money from the

8 16 September 2015 exploitation of labour, even if the good ef­ of our personal experience: when it be­ fects would outweigh the bad: it’s your life, comes embedded in the complex structure and it matters, morally speaking, what you of commitments, affinities and understand­ do with it, and not just – as Mac­Askill sug­ ings that comprise social life? We might gests – what is done because of it. even think that the arbitrariness of time That emphasis on ‘your’ is something that and place is transformed into something utilitarians often find conceptually mystify­ else, ethically speaking, through the ex­ ing, or at least a moral distraction. Here, change of a fleeting hug or thanks. What’s for example, is MacAskill talking about his more, MacAskill’s talk of fairness is too visit to the Hamlin Fistula Hospital in easy. It is no doubt unfair that some of ­Addis Ababa, and his later decision not to the world’s worst off are helped while oth­ donate to its main charitable benefactor: ers aren’t. But isn’t it just as unfair that the I’d hugged the women who suffered from this Ethiopian women MacAskill met are vic­ condition, and they’d thanked me for visiting tims of a debilitating condition that is too them. It had been an important experience costly to be ‘worth’ funding? And what of for me: a vivid first-hand demonstration of the victims of austerity or rising inequality the severity of the problems in the world. in the first world? MacAskill’s reminder This was a cause I had a personal connection that these people are still among the world’s with. Should I have donated to the Fistula richest is cold comfort (it also obscures Foundation, even knowing I could do more to help people if I donated elsewhere? I do what all those trampled by the ruling class not think so. If I were to give to the Fistula everywhere may have in common). Foundation rather than to charities I thought When MacAskill says that helping the were more effective, I would be privileging the Ethiopian women he met would be ‘arbi­ needs of some people over others for emo­ trary’ and ‘unfair’, he means to speak from tional rather than moral reasons. That would what the 19th-century utilitarian Henry Sidg­ be unfair to those I could have helped more. wick called ‘the point of view of the uni­ If I’d visited some other shelter in Ethiopia, or in any other country, I would have had a verse’. But in so doing MacAskill is try­ different set of personal connections. It was ing to step outside what is unavoidably the arbitrary that I’d seen this particular problem scene of ethical action: one’s own point of at close quarters. view. MacAskill thinks this self-transcend­ ence – or as close as we non-saints can get That word ‘arbitrary’ is striking. It is indeed to it – is essential if we are going to meet the arbitrary that MacAskill went to this hosp­ ethical demands of our day. Wittingly or ital and not another, in Ethiopia and not not, he believes, we are all like A&E doc­ some other country, just as it is arbitrary tors, forced to perform triage lest more that we have the family, friends, lovers and people suffer and die than have to. What is neighbours we do. But doesn’t such arbi­ required is impersonal, ruthless decision-­ trariness come to mean something else, making, heart firmly reined in by the head. ethically speaking, when it is constitutive This is not our everyday sense of the ethical

9 16 September 2015 life; such notions as responsibility, kind­ ness, dignity and moral sensitivity will have to be radically reimagined if they are to sur­ vive the scrutiny of the universal gaze. But why think this is the right way round? Per­ haps it is the universal gaze that cannot withstand our ethical scrutiny. There is a small paradox in the growth of effective altruism as a movement when it is so profoundly individualistic. Its utilitarian calculations presuppose that everyone else will continue to conduct business as usual; the world is a given, in which one can make careful, piecemeal interventions. The tacit assumption is that the individual, not the community, class or state, is the proper ob­ ject of moral theorising. There are benefits to thinking this way. If everything comes down to the marginal individual, then our ethical ambitions can be safely circumscrib­ ed; the philosopher is freed from the bur­ ­den of trying to understand the mess we’re in, or of proposing an alternative ­vision of how things could be. The philo­sopher is left to theorise only the auto­nomous man, the world a mere background for his righteous choices. You wouldn’t be blamed for hoping that philosophy has more to give. c

10 16 September 2015