Editor’s note: The first essay The Argument ever published asked, “How do we live with each other?” To mark our first anniversary on August 17, we’re returning to that founding question with a series about what we owe people whose beliefs and choices we may find difficult to accept or even understand. Over the next few weeks, familiar voices at The Argument and new voices alike will approach that question from different angles.
In today’s essay, Dylan Matthews notes that the AI boom is poised to unleash billions of dollars in philanthropic giving as it mints new millionaires and billionaires — many of whom have, let’s say, idiosyncratic views about the world. Is this private exercise of power a subversion of the democratic order? Is it an opportunity to try out bold new solutions to thorny problems? A little of both?
Personally, I think philanthropy is one of many productive release valves for a pluralistic society. Maybe I can’t get the government to spend tax revenue on public health in developing nations, but I can help do it with my own dollars. Or maybe you oppose abortion and are furious at the government for not providing other options for reluctant mothers; philanthropy lets you fund those if you wish.
Dylan, a one-time skeptic of big-time charities who is now a program officer at the major philanthropy Coefficient Giving, considers the dilemma of billionaire giving from first principles.
Disclosure: Coefficient Giving is an investor in The Argument.
To receive every essay in this series, subscribe today.
- Jerusalem Demsas
Back in 2018, the release in quick succession of works like Anand Giridharadas’ Winners Take All, Rob Reich’s Just Giving, and Edgar Villanueva’s Decolonizing Wealth led to a broad reconsideration in the commentariat of the ethics and political justifiability of large-scale philanthropy. I should know: I was part of it.
In 2019, I hosted a season of Vox’s podcast Future Perfect that focused on the relationship between philanthropy and democracy, particularly on the ways in which the former can undermine the latter. It wasn’t negative across the board — the episode on AI safety praising early funders of the field like Skype co-founder Jaan Tallinn has, I think, held up particularly well — but the main message was, “be suspicious.”
These concerns had been bubbling up for some time. But the broad anti-tech backlash that was surging due to concern over Facebook’s role in the 2016 election — and the growth of anti-billionaire politics fueled by both that and Bernie Sanders’ 2016 run — made 2018-19 the moment when these worries became mainstream.
I agreed broadly with Giridharadas, Reich, and Villanueva that philanthropy allows the wealthy to exert excessive, and unjustifiable, influence over society; that the charitable tax deduction unjustly forces ordinary Americans to fund that influence; and that the unaccountable nature of foundations, not constrained either by voters (as governments are) or market pressures (as businesses are), made them inherently dangerous, sometimes to the point of enabling outright atrocities.
But the 2010s era of Big Philanthropy is going to look tiny compared to the giving unleashed by the impending Anthropic and OpenAI IPOs. The OpenAI Foundation’s shares in OpenAI are worth, as Stripe executive Nan Ransohoff has noted, around $220 billion already, larger (and much more poorly diversified) than any other foundation endowment on earth. All of Anthropic’s co-founders have committed to giving away 80% of their net worth, which, given Anthropic’s current valuation, should mean another hundred billion or more in giving.
Close observers have known this was going to happen for a while, but after Ransohoff’s essay, the impending flood of AI money has become common knowledge, sparking coverage from The New York Times to The Economist to the Financial Times. And with that attention has come some familiar debates about whether this level of giving is acceptable in a free society, or whether it’s, in David Wallace-Wells’ words, “an assertion of plutocratic control over yet more aspects of collective life.”
There was and remains plenty to be suspicious of in the past and present of big philanthropy. But my views have evolved a bit since 2019, which is perhaps unsurprising as during the same period I left journalism to join Coefficient Giving. Coefficient Giving (formerly “Open Philanthropy”) is a large philanthropic advisor and grantmaker that supports work in areas like animal welfare, global health, pandemic preparedness and prevention, and AI safety. Last year we recommended some $1 billion in grants; we’re on track to significantly increase that number this year, and our advisory work in particular has grown considerably as the AI rich seek to give away their new fortunes.
I wouldn’t work where I do if I didn’t think our project is compatible with, and indeed beneficial to, a liberal democratic society. More importantly, though, I think the events of the subsequent seven years, and specifically of the past two years, have made the most scathing version of this critique less tenable than ever.
Implicit in the critique that philanthropy conflicts with democracy is the notion that decisions about the distribution of wealth and power should be made principally by the government, not by private actors. I believed this to be true in 2019. But now I think this is mistaken — and, indeed, somewhat illiberal. The effects of the second Trump term on foreign aid and science are the clearest indication of the limits of this approach.
Rather, no single entity, including the government, should monopolize these decisions. Foundations and grantmakers make serious mistakes. But the important thing is not that grantmakers are perfect, but that our errors are different from those made by the state and private business. Multiple imperfect institutions making distinct errors together add up to a more stable and resilient overall system than privileging one approach over all others.
A role for foundations
In the field of risk assessment, analysts often talk about “Swiss cheese models.” The idea is that in a complex system, each step in the process will have natural failures. But if you layer multiple steps with different failure points over each other, you can emerge with a system that’s robust to each of those failures, the same way a block of 10 Swiss cheese slices does not have a single hole going all the way through.
You can think of the way we provide goods in American and global society as several different, and differently flawed, institutions overlapping: markets, governments, religious institutions, schools and universities, and philanthropy, to name some of the most important. If we’re lucky and/or politically wise, these institutions will complement each other well and help remedy each other’s flaws.
The failures of philanthropy as a model are clear. It over-privileges the views of the wealthy. It creates power-laden relationships that, even in my short time as a grantmaker, I have already at times found awkward and uncomfortable (and the grantmaker’s discomfort is probably the least of the issues there).
But the case for philanthropy isn’t that it’s perfect. It’s that philanthropy lacks certain key defects of its complementary institutions, like markets and government.
