top of page

"Effective Altruism, Longtermism, and the Problem of Arbitrary Power" by Gwilym David Blunt


White house on hill

From The Philosopher, vol. 111, no. 2 ("Where is Public Philosophy Going?").

If you enjoy reading this, please consider becoming a patron or making a small donation.

We are unfunded and your support is greatly appreciated.


Impact is a word that gets bandied around a lot in academia. It is the metric by which success is measured and managers appeased, but for anyone working in philosophy it can appear almost impossible for one’s work to reach outside of the small network of professional academics and passionate students. Philosophy seems trapped in the ivory tower. That is what makes William MacAskill exceptional. His work as a philosopher has had a real impact outside of academia, not once but twice.


First there was effective altruism. Inspired by Peter Singer’s work on ethics, its most visible advocates are MacAskill and Toby Ord. Effective altruists argue that not only do people have an obligation to ‘do good’ unconstrained by the borders of their states, but they must also do good in the most effective way possible. You may love stamp collecting, for example, and would feel extraordinarily happy if you gave a £1 million to the Royal Philatelic Society London. Yet, with apologies to the philatelists reading this, effective altruism would say that this would be the wrong way to use your money; it would be unethical. Instead, you should take your money and give it to an organisation that has a strong track record in providing treatment for river-blindness in Sub-Saharan Africa.


This is very commonsensical and helps explain effective altruism’s transformation into a major social movement. From its origins in Oxford, it has spread across the world, and effective altruist-aligned organisations, such as MacAskill’s Centre for Effective Altruism, have annual expenditures approaching $400 million with an estimated $46 billion more in funding commitments. That is serious impact.


Then came longtermism. Whereas effective altruism extends the spatial horizons of our ethical obligations beyond borders, longtermism extends the temporal horizons of these obligations to future people. It posits that a major ethical priority, perhaps even the major ethical priority, is to circumvent ‘existential risks’ to humanity. These threats range from planet-killing asteroids to nuclear war to rogue artificial intelligence going ‘full Skynet’ and annihilating humanity. Some are incipient, others may be some way off. Even if some of the risks seem a bit fantastical, it is easy to see the appeal of longtermism, given that human extinction or a barely better dystopian future does not seem particularly inviting, especially when the prospect of preventing such a future is plausible.


Longtermism has built on the success of effective altruism. Both Ord and MacAskill have published books on it which have gained significant media attention, both sold an enviable number of copies for philosophy, and MacAskill’s What We Owe the Future even gained a public ‘attaboy’ from Elon Musk for being ‘a close match to my philosophy’. Again, this is the sort of impact that makes vice-chancellors delirious with joy.


Yet, while effective altruists and longtermists were dizzy with success and proceeding with a reckless innocence, they failed to truly appreciate the power they were accruing. In the summer of 2022, MacAskill did an interview with The New Yorker in which he expressed anxiety about the shift in the culture of effective altruism from the ascetic frugality of students to the business culture of a well-funded multinational organisation.


The drift from austerity is best embodied by Effective Venture Foundations, of which MacAskill is a trustee, purchasing Whytham Abbey in Oxfordshire as its base of operations thanks to a £17 million grant from Open Philanthropy. Critics, inside and outside the movement, claimed that it was hypocritical to spend millions on an Elizabethan manor house rather than use it to improve the condition of the world’s poorest people. The rebuttal is that a proper headquarters is necessary to build the movement, generate innovative ideas, and make a better impact in the future. Putting it plainly, monkish austerity isn’t a great selling point; if you can ‘do good better’ with a mansion, then aren’t you ethically obliged to have one? This is why MacAskill viewed this as merely an ‘aesthetic’ issue.


Things may have turned out differently if a copy of Philip Pettit’s Republicanism was sitting on their bookshelves beside Peter Singer’s The Life You Can Save to caution effective altruism and longtermism about the hazards of power beyond questions of mere efficiency. But they didn’t and things went wrong.


***


The difficulty of breaking out of the campus and MacAskill’s undeniable success might explain the atmosphere of schadenfreude among other philosophers when the cryptocurrency exchange FTX collapsed under accusations of fraud and criminal mismanagement. Unfortunately, Samuel Bankman-Fried, the CEO of FTX, was an outspoken advocate of effective altruism and longtermism with longstanding personal ties to MacAskill. Bankman-Fried previously worked at the Centre for Effective Altruism, while MacAskill was an advisor for FTX’s philanthropic-wing, The Future Fund. This fund committed millions of dollars to the Centre for Effective Altruism and other organisations with which MacAskill is or was affiliated. The U.S. Security Exchange Commission has accused Bankman-Fried of defrauding investors of $1.8 billion. That’s not the sort of impact anyone wants.


Overnight it seemed like the public soured on effective altruism and longtermism. The previously friendly mediascape turned hostile with critical pieces appearing in The New Yorker, Time, and The Guardian. The thread running through these critiques was that effective altruism (and longtermism by extension) was complicit in Bankman-Fried’s alleged crimes, that financial malfeasance can be overlooked if the result, such as the $160 million committed by the Future Fund to various causes, produces a good outcome. This has been broadened by critics like Alice Crary to argue that effective altruism and longtermism have a dirty hands problem with the ruthless and destructive world of capitalism. Whatever good effective altruism does cannot offset the harm caused by a system that facilitates sufficient capital accumulation to become an effective altruist in the first place.


Power for effective altruism and longtermism is mostly a benign tool that can be used to help people. It should be used effectively (with a side-constraint against causing harm). And that’s about it.

Effective altruists came out swinging in its defence. MacAskill took to Twitter and pointed out that in What We Owe the Future he said that there was an obligation to do no harm. He also expressed his ‘utter rage’ against Samuel Bankman-Fried for the harm he caused and ‘self-hatred’ for being deceived. Peter Singer joined the defence, arguing that while deception is sometimes justified, such as lying about the location of Jewish fugitives to their Nazi persecutors, there was no such justification in the case of Samuel Bankman-Fried. It was ‘unnecessary and unjustified’. This defence admits that they were duped by an unethical huckster, but denies that there is any serious flaw in the movement itself. It is simply an unhappy coincidence that Samuel Bankman-Fried was associated with them.


The reality is not so simple.


The FTX fiasco reveals a problem deeper than keeping bad company and more subtle than anticapitalism. It exposes a naivety about power, or the consequences of, as Peter Wolfendale put it in his review of What We Owe The Future, the absence of ‘a working theory of power’. This naivety has allowed a rot to set into effective altruism and longtermism. It has made effective altruism too comfortable with uncontrolled power and too willing to internalise the worldview of their ultra-wealthy patrons.


Power for effective altruism and longtermism is mostly a benign tool that can be used to help people. It should be used effectively (with a side-constraint against causing harm). And that’s about it.


There does not seem to be a difference between the power one uses to rescue the drowning child in front of you or the exercise of power in a pharmaceutical company’s research programme or the well-run NGO helping people out of poverty. There is little engagement with the complexities of institutionalisation or the entanglements of knowledge and power. It is a concept of power that sits comfortably with the privileged. It explains effective altruism’s popularity among the students and graduates of elite universities, to say nothing of the billionaire class. These are people who have benefitted from the exercise of power and have little reason to interrogate it too closely. It is also a conception of power that explains the apparent shock at Bankman-Fried’s actions.


Singer and MacAskill have reacted to this scandal as a failure of personal morality. The failure is about Bankman-Fried’s character rather than the structure of effective altruism as a social movement or as a form of ethics. However, seeds of this problem were evident long before Samuel Bankman-Fried brought effective altruism into disrepute. In 2015, I wrote that Peter Singer’s proto-effective altruism was poorly equipped to deal with the institutions and power necessary to realise duties to impoverished distant strangers; it was too focussed on results and neglected how even benevolent power can produce problems of justice. This is a foundational problem for effective altruism and one that MacAskill intuited in his New Yorker interview but has not addressed.


Effective altruism, when it began, had ascetic norms designed to constrain power, but slowly and surely they have been eroded by arguments about ‘effectiveness’.

The point here isn’t to criticise the large amount of capital that effective altruist organisations have diverted into internal growth; perhaps they are correct and owning manor houses makes their organisations more effective. Instead, it is to identify a stress point in effective altruism and longtermism: they are pitched as personal ethics, as actions that you can take one your own, but require complex forms of institutionalisation and the exercise of power to achieve their ends. If you read most effective altruist literature, it is pitched at the individual reader (“What good can you do?”). When it comes to questions like “What institutions should you support?” the answer is merely “the most effective”. MacAskill misunderstands his own anxiety; this isn’t about the optics of giving, but the constraints we put on the exercise of power. Effective altruism, when it began, had ascetic norms designed to constrain power, but slowly and surely they have been eroded by arguments about ‘effectiveness’, and as they receded the absence of strong institutional constraints became apparent. This leaves effective altruism and longtermism exposed to what we may call ‘the despotism trap’.


***


The constraints we place on power are often portrayed as obstacles rather than safeguards. There is a certain wisdom to this, especially if we consider someone like Bill Gates and the work of the Bill and Melinda Gates Foundation. Although Gates has no formal connection to effective altruism, Singer has called him and Warren Buffet the ‘greatest effective altruists in human history’. The Gates Foundation is one of the world’s most significant NGOs, with an endowment approaching $50 billion and every year it spends roughly $5 billion to fund programmes in education, development, and global health. This makes it comparable to a major developed state such as Australia in terms of foreign aid spending, but it tends to be more effective and influential.


This is especially true in the case of global health, where the Gates Foundation has played a key role in setting the health agenda and providing far-sighted guidance. Long before Covid-19, Gates was actively warning that we were unprepared for a global pandemic, helped to set up the Coalition for Epidemic Preparedness Innovations, and funded vaccine research. Indeed, during the pandemic when President Trump withdrew the United States from the World Health Organisation, the Gates Foundation was in line to become its biggest funder. Without the constraints that affect the state, the Gates Foundation has been more farsighted and innovative in anticipating and preparing for long-term threats to humanity.


The Gates Foundation, if we disassociate it from the dirty hands problem that enables someone to become a billionaire and set aside controversial moments like Gates’ initial opposition to lifting patents on Covid vaccines, does good work and is very effective. Perhaps the occasional Samuel Bankman-Fried is tolerable if it enables a Bill Gates to improve the lives of millions if not billions of people? We can look at a thought experiment I like to call the Indecent Proposal to see why this is not the case.


The existential risks that concern longtermism appear to overwhelm any concerns about constraining power.

Let’s say you are at home one day and there is a knock on your door. And who is it? A billionaire and he has an offer for you. He will give you a better lifestyle than you could reasonably achieve under ordinary circumstances, but there is a small catch. You must become his slave and as such he would have uncontrolled power over your life. However, he promises that he would never actually use this power (or at least he would never use it in a way that wouldn’t be to your ultimate benefit). The billionaire is thought of as a pretty decent person, he can occasionally seem eccentric but has done a lot of philanthropic work. Would you accept this offer?


I know that I wouldn’t. No matter how tempting the lifestyle would be, slavery seems obviously wrong. I will confess to smuggling in this premise: human beings are, generally speaking, agents capable of determining a concept of a good life and pursuing it with a degree of rationality. This capacity for autonomy is something that requires respect and protection. That is to say they have rights, including the right to some meaningful control over the institutions that have a profound impact on their lives. If your rights rest on someone else’s whims, you don’t have a right, you have a privilege. And privileges can be taken away in an instant. It doesn’t matter whether the person who holds unaccountable power is benevolent or not, a gilded cage is still a cage.


***


This is magnified when it comes to longtermism and existential risks. Alice Crary has pointed to the willingness of longtermists to give existential threats to humanity such weight that they deprioritise actual suffering in the world that we live. However, the problem is deeper. The existential risks that concern longtermism appear to overwhelm any concerns about constraining power. Existential risks parallel the ‘supreme emergency’ that dictators so often use to justify the destruction of the rule of law and democratic institutions. The “only I can fix it” mentality seems equally at home in the mouth of authoritarians and longtermists.


Uncontrolled power, or domination, is the central concern of republican political philosophers, like myself, but it is difficult to see how anyone who values basic human autonomy would not be concerned about it. When power can affect people’s ability to live their lives, it needs to be placed under constraints, even at the cost of its effectiveness. This seems almost self-evident when we discuss politics. The above defence of the Gates Foundation as a vehicle for effective altruism is strikingly similar to how authoritarians are credited with being more effective than constitutional democracies; they ‘make the trains run on time’ and so forth. Yet, most political philosophers would not say that the costs of a Caligula are tolerable if you also get a Marcus Aurelius. We want institutions set up to prevent Caligulas and that will also enable a good ruler to still operate effectively.


The ethos of Silicon Valley is often described as ‘libertarian’, but it isn’t. It is deeply authoritarian. It worships the ‘great genius’ – whether it’s Elon Musk, Samuel Bankman-Fried, or Bill Gates.

The constraints found in MacAskill and others are minimal. The most prominent is the prohibition against harm. This is essentially a Hippocratic Oath for giving. It is better than nothing, but far from good enough. The Hippocratic Oath doesn’t mean much if there isn’t also professional association and external oversight to ensure that the oath is kept. This is comparable to the ‘slave codes’ in the antebellum United States which prohibited masters from excessive cruelty to their slaves, but were unenforceable because slaves lacked the status to hold their owners to account; they existed more as property than as people under the law. These codes rested entirely on the good conscience of slave-owners, something that we might say is intrinsically compromised based on their ownership of other human beings.


There is an irony here because one of the concerns of longtermism is ‘value lock-in’. They worry that a nascent artificial general intelligence will be programmed with illiberal values that could produce a stagnant or dystopian future for humanity. Yet, effective altruists and longtermist cannot see how their own movements are in danger of having illiberal values locked into them. Marx wrote that the point of philosophy is to change the world, but sometimes the world changes philosophy.


***


The intimacy between effective altruism, longtermism, and Silicon Valley has been much noted. The ethos of Silicon Valley is often described as ‘libertarian’, but it isn’t. It is deeply authoritarian. It worships the ‘great genius’ – whether it’s Elon Musk, Samuel Bankman-Fried, or Bill Gates. These people are seen as epistemically privileged; they are innovators who can see problems and solutions that elude us ordinary folk. It rejects constraints against these visionaries as the pettiness of small minds. They drape themselves in the regalia of liberty to challenge government, but their interest is in being able to exercise uncontrolled power. The excuse is that the ends will justify the means. We can see how these attitudes have influenced effective altruism when it overlaps with ‘philanthrocapitalism’, whereby people like Gates bring the mechanisms of the market into philanthropy on the grounds that is more effective. Effective altruism seems willing to play with the notion that billionaires are epistemically special, that the marketisation of giving is the most effective path, and that constraints should be avoided for the greater good. If this Silicon Valley authoritarianism continues to shape the epistemic horizons of effective altruism and longtermism, they risk becoming ‘a mirror for tech-bros’.


What effective altruism and longtermism need is a theory of power that can appreciate the risks of institutionalisation and threat of arbitrary control, epistemic or otherwise.

Yet, I do not want to join the pile on. Effective altruism has inspired a lot of people to try to make the world a better place; it deserves respect. Likewise the schadenfreude around MacAskill’s difficulties is distasteful. What effective altruism and longtermism need is a theory of power that can appreciate the risks of institutionalisation and threat of arbitrary control, epistemic or otherwise. Some effective altruists have recognised this danger. Carla Zoe Kremer, for example, has proposed that there ought to be greater democratisation within the power structure of Effective Altruist organisations, that there should be whistle-blower protection, and internal checks and balances. Yet, at the time of writing, these sensible constraints have not been adopted. They are still caught in the despotism trap. This makes both of these movements vulnerable to people like Samuel Bankman-Fried, but the risks are still greater. They are locking in institutional values that run against basic human autonomy by granting uncontrolled power to those who are perceived to have good intentions. We owe the future more than that.


Gwilym David Blunt is a Lecturer in International Relations at The University of Sydney, a Fellow of the Ethics Centre, and Senior Research Fellow of the Centre for International Policy Studies. His research examines freedom, inequality, resistance, and global distributive justice. He is the author of Global Poverty, Injustice, and Resistance (2019). Website: https://www.gdblunt.com

 

From The Philosopher, vol. 111, no. 2 ("Where is Public Philosophy Going?").

If you enjoyed reading this, please consider becoming a patron or making a small donation.

We are unfunded and your support is greatly appreciated.

Comentários


bottom of page