Getting it right (Part 4: Criticism)

Group discussion is not always efficient. When ­people have their ideas closely aligned to start with, it leads to polarization. When ­people start with conflicting ideas and no common goal, it tends to exacerbate differences. Group discussion is typically beneficial when participants have differ­ent ideas and a common goal

Hugo Mercier and Dan Sperber, The enigma of reason

1. Introduction

This is Part 4 of my series Getting it right. This series highlights ways in which effective altruists think and act admirably.

Part 1 discussed altruism. Many effective altruists donate large parts of their income to charity, and some go further by risking their health in challenge trials or even donating organs to total strangers.

Part 2 discussed global health. Effective altruists were instrumental in the push towards more rigorous and evidence-based methods in global health, and have committed well over a billion dollars to global health causes.

Part 3 discussed ambition. Effective altruists dream big, and they are often successful in bringing about large-scale change.

Today’s post discusses criticism. It is hard for movements to engage with their critics. While no one is perfect, effective altruists often do a remarkably good job engaging with their critics.

2. Why engage with critics?

An increasingly popular view in cognitive science holds that reasoning is for argument. On this view, humans often reason with the goal of argumentatively supporting beliefs that they or their communities already accept.

This view predicts that individual reasoning should show some pathologies. It should show limited accuracy, because reasoning can be directed at supporting existing beliefs even when those beliefs are not true. Individual reasoning should show self-serving biases, giving rise to a psychological immune system in which reasoning is used to help us to think better of ourselves than the evidence supports. And individual reasoning should show a range of myside biases, such as confirmation bias, which favor an individual’s own side in an argument.

Suppose, however, that we are not so locked inside our own perspective that we cannot recognize the truth when we see it. Or, barring that, suppose that the spectators to an argument can at least do this, even when the reasoners themselves cannot. This suggests that reasoning will often perform well in diverse groups. Someone within the group may stumble upon and advocate for the correct answer. In the process of advocating for this answer, they will find the truth. And most of the rest of us will recognize the truth as arguments for the truth win out.

However, for group deliberation to work well, two things are necessary. First, groups must be diverse. A group of likeminded individuals constitutes an echo chamber, and is likely to reason itself into ever more confident and extreme views. Adversarial reasoning cannot work without an adversary.

Second, reasoning must be adversarial. This does not mean that we should treat one another as enemies, or even speak unkindly. It means rather that people with diverse opinions need space to publicly disagree with one another. If the truth cannot be voiced, it will not be found by the group.

If this is right, then argumentative engagement with critics is essential to the effective altruist project. Engagement should increase accuracy and decrease biases. Lack of engagement risks polarization, bias, and confident extremism.

For this to work, engagement must not be shallow or nominal, but sustained and genuine. That is not an easy thing to do. Most groups do not like to listen to extended discussions about why they are wrong. But effective altruists often do this, and that is to their credit.

3. Engagement with critics

Effective altruists actively read and engage with their critics. Many, if not most of my own readers are effective altruists. My own work has been discussed in at least 115 posts on the EA Forum, including posts discussing many of my papers and several critical evaluations of my arguments. I am often invited to speak at effective altruist events, listened to, and approached afterwards for follow-up discussions.

More broadly, effective altruists have extensively engaged with criticism including:

The good it promises, the harm it does: Critical essays on effective altruism: Reading group, review, review.

Zoe Cremer and Luke Kemp: Here.

Émile Torres: Here, here, here.

Eric Schwitzgebel: Here.

Just to name a few.

This is a good thing. Engagement with critics promotes diverse and adversarial deliberation of exactly the sort needed to avoid the worst effects of homogenous deliberating groups.

4. Encouraging production of criticism

Engaging with critics is already admirable. But effective altruists do not stop there. They often actively promote criticism of their views.

In 2022, the EA Forum ran a ($120,000) contest for criticism and red-teaming of leading views.

In 2023, Open Philanthropy ran a ($225,000) AI Worldviews Contest focused on criticism of its AI worldviews.

There is an entire podcast dedicated to criticism of effective altruism.

This blog was previously funded by a grant from Manifund. (In fact, they awarded me twice what I asked for within a day of my initial ask). My research is currently funded by the Survival and Flourishing Fund, and I have previously received grants from several EA-aligned funders.

There are very few groups in the world that do this much to promote criticism of their own views. I do not do half as much to promote criticism of my own views.

5. Self-criticism

Another way for groups to generate criticism is for group members to criticize the group themselves. There have been strong and persistent internal critiques of effective altruism, which have been often heeded and engaged with.

In 2023, a group of Concerned EAs launched a detailed debate concerning a number of challenges facing the EA community.

Many prominent effective altruists have critically and publicly reflected on developments in effective altruism.

Effective altruists are notably unafraid to target particular works which they view as subpar. For example, Yudkowsky and Soares’ recent book was critically reviewed by Will MacAskill and in Asterisk Magazine.

Not all of these criticisms have been favorably received. Many critics feel that they have not been treated exactly as they would like. But by comparison to other communities, internal critics of effective altruism are surprisingly numerous and well-treated.

6. Taking stock

Engaging with critics is hard. They tell us that we are wrong, and that we are behaving badly. It is doubly difficult when those critics come from within our own group.

Effective altruists make a concerted effort not only to engage with their critics, but also to encourage the production of criticism, spending their own money if necessary to make sure that high-quality criticism is produced.

That is a very healthy epistemic practice that does a lot to fend off troubling dynamics seen in homogenous deliberative groups.