Why & How to Make Progress on Diversity & Inclusion in EA

This post is a collection of potential solutions I’ve come to over 3+ years of observing, experiencing, thinking about, reading material relevant to, and discussing issues of diversity and inclusion in the EA community, assisted by my experience in other communities.

“Diversity” is about representing people from diverse walks of life. “Inclusion” is somewhat more nebulous, and there seems to be a misunderstanding about its meaning in the community: Inclusion is not about welcoming everyone in, it’s about welcoming in the right people and ensuring we’re not excluding them for irrelevant criteria. I think the people who the effective altruism community should work to engage — and I assume this isn’t very controversial — are people who want to do the most good, or at least people who are interested in doing good better.

There has been a lot of loose discussion of these issues, mostly from those directly affected, but few actions taken to seriously address them. This post will address gender-based exclusion more than other issues as that’s the one I have the most knowledge on, but a lot of the practices I suggest should make the community more inclusive of a broader diversity of people. My goal is to keep the ball rolling by spurring further discussion on solutions and helping people implement the most promising ones, so especially if you are from an underrepresented group and/​or have expertise in this area, please do comment with your own thoughts, information, and ideas. Feel free to message me and I can post your comment anonymously if you prefer.

Why is this something we should pay attention to?

Most people in the EA community who I speak with agree this is an important issue, but for those who don’t, I’d like to formally lay out the reasoning in this section.

Based on our demographics, my observations, and many conversations with women, people from other underrepresented backgrounds, and even people from overrepresented backgrounds who still felt or feel the community is too exclusionary, I think the EA community is not quite selecting for “people who want to do the most good” or the lighter version of that, but people who are both that and young, white, cis-male, upper middle class, from men-dominated fields, technology-focused, status-driven, with a propensity for chest-beating, overconfidence, narrow-picture thinking/​micro-optimization, and discomfort with emotions. These features suggest limitations of our capabilities, both individual and collective, that could be relieved if we worked harder on diversity and inclusion.

I’ve met many people who are deeply driven to help others as much as possible and who beyond that are highly capable — e.g. high analytical ability, years of experience with nonprofit management, other specialized skills, graduate degrees in relevant fields — but who left, limit their involvement in, or never joined the EA community because of their experience with its culture and norms. Some of those who have stuck around do so begrudgingly because the community still offers them enough to be worth it or because they think the community has so much potential that bearing it to contribute what they can is worth it, but they’re not giving us all they have to offer, and many others are turning away entirely. One big effect here seems to be the exclusion of women, as suggested through my conversations with many women and the community’s gender ratio of roughly 70% men and 2.7:1 men:women. The exclusion of people of color is a noticeable problem as well, with e.g. black and hispanic persons severely underrepresented compared to U.S. demographics. We’re losing the potentially huge amounts of resources that such people could bring to the EA movement: knowledge, experience, management ability, perspective, ideas, creativity, analytical ability, emotional understanding, social competence, big-picture thinking, enthusiasm, career capital, career opportunity, a variety of specialized skills, networks, money — you name it.

See also Alexander Gordon-Brown’s post on some other characteristics EA is missing out on in terms of diversities of talent, experience, opinion, and appearance.

Not only are we missing out on those individuals and their resources themselves, but a sum that would be greater than the whole of its parts: [Edit: As a commenter noted, the content of following sentence is debated in psychology.] A group’s collective intelligence is only moderately related to its individuals’ intelligences, and gender-diverse teams score higher on collective intelligence than all-male or all-female teams (What Works: Gender Equality by Design, 10). Research also shows that diverse teams are more creative, more innovative, better at problem-solving, and better at decision-making — see Georgia Ray’s post “Diversity and team performance: What the research says” for more detail. Companies in the top quartile for diversity in gender and ethnicity are 15% and 35% more likely to outperform their industry’s median performance, respectively, and companies in the bottom quartile lag behind the median. Fortune’s top 50 workplaces for diversity list an average 24% higher year-over-year revenue growth than companies that didn’t make the list, and companies with multiple women in the C-Suite are more profitable. These are all correlations, but the effect sizes are very large and the causal explanation seems highly plausible.

There are also known issues in EA that seem likely to be mitigated or eliminated through an increase in diversity. For example, this year’s EA Global San Francisco conference focused on shifting the community towards “doing good together.” Women tend to be more collaborative than men, so if the community had better gender representation, we could already be thinking big and emphasizing doing good together.

Even if people in our community are less prejudiced than the rest of society, small biases can have big impacts: One simulation found that bias accounting for only 1% of variance in evaluation scores resulted in the top level of the simulated workforce being only 35% comprised of the discriminated-against group, instead of 50% like the original pool (What Works: Gender Equality by Design, 14).

Unfortunately I suspect some people in the community are content, implicitly or explicitly, to assume that women and people of color are inherently so much worse than white men at thinking about altruism effectively that the constitution of the community is merely an effect of this presumed difference, and that as such putting effort into diversity and inclusion would either be too difficult and costly to be worthwhile or would dilute the community. I find this argument lacking given the alignment of that thinking with demonstrated biases in society at large — i.e. people tend to think that women are more intuitively-driven and less analytical than men, which does not seem to be borne out and in fact the opposite may be more likely — and given the suspiciously large gender and race disparity in EA, as well as the the very small size of the community at present. The latter enables us to target selectively, not randomly from the general population, even if this loaded, simplistic, and to my knowledge unfounded claim is true. Moreover, there are many examples of women, white and of color, who felt or feel excluded by the EA community, despite being entirely onboard with the philosophy of EA, having participated in the community for years, and having made major changes in their thinking and lives because of EA — they just really dislike the community.

Relatedly, some may assume that our community is genuinely merit-based — that we simply reach out to and include the most qualified people, regardless of their race, gender, etc. Did you know that, at least in an experimental setting, when organizations espouse meritocracy managers show greater gender-based discrimination than those at other companies? And that having a gender quota is more likely, assuming probably that an organization is competent at hiring, to weed out mediocre men than to introduce mediocre women? Unfortunately most of the specific details of the conclusive and pervasive sexism I have experienced, seen, and heard of first-hand within the EA community are confidential — and I don’t just mean sexual harassment and assault, there are other more pernicious and more prevalent forms of sexism in society and in the community, such as the holding of women to higher standards of competence and the consistent underestimation of women’s trustworthiness (What Works: Gender Equality by Design, 27). Happily some of it has been acknowledged by its offenders, who have in some cases stated credible intentions to improve — though whether they are in fact improving takes time to assess, and without ongoing personal or even cultural support they may find improving difficult. Even without the details of specific experiences people have had, it should be sufficient to observe that there are many examples of qualified people being excluded, and no evidence has been offered to justify the assumption — which was recently voiced, to no opposition as I understand it, on a panel at EAGxBerlin — that we are merit-based. [Edit: I want to note as well that who EA seems to select for matches exceptionally well with privilege in society at large, which would be quite a coincidence.]

Some people in the community have made other thoroughly unreasonable claims to justify the status quo, such as that women would be a distraction in the workplace. If they are, the problem is entirely the men who can’t adhere to basic professional norms and who presume their contributions so important that the minor cost to them of being less sexist outweighs all of the potential contributions of all the women they’re keeping out. To my knowledge this claim was recanted — under pressure or reflection, I’m not sure — but it’s a red flag for other sexism. Someone else has said women aren’t as willing as men to take low salaries for altruistic purposes, apparently in ignorance of the rest of the nonprofit world, whose volunteers and workforce are overwhelmingly women. Such unrigorousness should be thoroughly discouraged.

I think the majority of the problem, however, is that while many people know we have a problem, they don’t know what they themselves can do about it.

What changes can we make to more effectively select for the right people?

The evidence base on effective strategies to reduce prejudice and increase inclusion in general is weak, though growing. I don’t claim that the following are all of the answers, nor necessarily the best answers, nor even that they’re all right or involve no tradeoffs. My aim is just to put my ideas out there in the interest of continued discussion and action on this issue. I also don’t claim to be perfect in implementing these myself, but I do generally aspire to embody those I’m failing in.

● Recognize that there is a problem, in society at large, in the communities EA sources from, and within the EA community. Even if you are not convinced by the evidence I’ve presented about why this is a problem our community needs to address, you should still be compelled by the fact that so many people both in EA and elsewhere think we have a serious problem. Look to data — I include barely any of the literature on sexism and other systematic biases in this post because it is vast and Googleable — and to accounts of people in the community — or no longer are — who are from the groups in question. Do not rely on your intuitions or those of anyone lacking the perspectives of people from underrepresented groups. If you disagree that this is an important problem or about any of the steps I suggest to make headway on it, let’s have a discussion so we can get to the truth of the matter.

● Recognize that it is extremely probable that you harbor biases that you are not accounting for. Recognize that recognizing bias in society and our community isn’t enough — people tend to think they are less biased than average, and tend to demonstrate the same levels of bias even when they are experienced with seeing a bias, made explicitly aware of the bias, and asked to introspect to ensure they are not making a biased judgement (What Works: Gender Equality By Design, 45-48). The latter can even backfire, which may be the effect of this whole statement, but I think transparency is sufficiently important to outweigh that risk. Even if you have evidence that you are successfully debiased in some ways — e.g. calibrating against overconfidence in online tests — the society you grew up in has many biases, and you are highly unlikely to be exempt from all of them. People in the EA community might even be particularly susceptible to some.

● Don’t penalize the “heart” as though there is only the “head.” EA is both, and one is nothing without the other in this movement. I prefer to play the long game with my own investments in community building, and would rather for instance invest in someone reasonably sharp who has a track record of altruism and expresses interest in helping others most effectively than in someone even sharper who reasoned their way into EA and consumed all the jargon but has never really given anything up for other people. I see exceptions to this being the best investment on the whole, but none who I think wouldn’t be here anyways if we were focusing much more on the former personalities. In practice, the lowest-hanging fruit to elevate the heart is to be empathetic with and kind to people. At the very least, ensure you are not being dismissive of people’s emotions, and in particular feminine-coded emotions like empathy, grief, sadness, or love — things that drive a lot of people’s altruism. Some of the most talented and resolute people in this community are here because they are deeply emotionally compelled to help others as much as possible, and we’re currently missing out on many such people by being so cold and calculating. There are ways to be warm and calculating! I can think of a few people in the community who manage this well.

● Recruit and promote women to manage teams. Women tend to be better managers than men.

[Edit: Additional suggestion. People in high places in the movement, particularly white men, publicly state the importance of EA being diverse and inclusive to you.]

● CEA and EAF could both, or jointly, hire a Diversity & Inclusion Officer. CEA and EAF, your intention is to be institutional leaders of the EA community, so lead the way on this critical aspect of movement-building — there is definitely a full-time job’s worth of advising and other work to do, probably even just with the suggestions I list here. Some universities and companies have such a position, and I — and I’m sure others — would be happy to advise on what the position’s responsibilities would look like. (Thank you Sana Al Badri for this suggestion.)

All organizations should hire communications staff who are versed in inclusionary communications practices. Alternatively, the Diversity & Inclusion Officer could train them.

● Adopt and enforce a clear policy — as organizations and individuals — for dealing seriously and fully with illegal actions like sexual harassment and explicit discrimination or discrimination revealed by HR or legal counsel. Commensurate consequences and reform procedures, escalating as necessary to expulsion, are critical. The perpetrator is not so much more important than the greater number of people they are driving away, the risk of a lawsuit to the organization protecting them, or the risk they bring to the community’s reputation, that such actions should be protected. If this community’s members are as smart as we like to think, using a heavy hand once, if necessary at all, should be all it takes, so long as the threat of using it again is credibly maintained.

● If you go out with colleagues, ensure you’re not just including the ones most like you. A lot of opportunity to build skills, network, and advance one’s career happens out of the office, and favoring some colleagues over others can lead to systematic disempowerment. If you are a man and can’t go out with women colleagues without thinking of them sexually and making the interaction uncomfortable, or if you can’t have a conversation about work and EA with women colleagues at lunch, you should not be managing anyone.

● If you see something, say something. Don’t leave the reporting of problematic behavior to the people who directly experience it. They are feeling disempowered and alienated and are usually in a far less capable position to do something about it.

● If you experience something, try to at least say something to someone. Whether you decide it is in your interest to say something or not, ensure you at least consider the risks to other people and the broader community if you do not. I appreciate that in many if not most cases we just want to move on with our lives, and this burden should, as noted above, not be left to the people experiencing the problem, who generally face higher risk bringing it up than other people.

● We could establish a website providing resources for legal counsel and enabling people to anonymously share experiences regarding discrimination, harassment, and assault, both to inform less-aware community members of issues in the community and to provide a sense of accountability to the movement as a whole, as the testimonies would be publicly accessible.

● Update your valuations of men’s competencies downwards, and of women’s upwards, particularly when you are forming your first impressions. People already inaccurately perceive women as less competent than men, even when their work is superior, in addition to which men overestimate and oversell themselves while women underestimate and undersell themselves. Yes, this will penalize the rare men who represent themselves accurately or under-represent themselves, and favor the rare women who represent themselves accurately or over-represent themselves, so take care, but the risk of overcorrection is not sufficient reason to resort to the prejudiced status quo. Additionally, in more long-term and formal environments, utilize standardized and objective metrics of competency whenever possible, such as trial projects when hiring. Relatedly, consider promotions and do hiring in rounds, not on a rolling individual basis.

● Amplify the contributions of people from underrepresented groups, in personal interactions, meetings, articles, podcasts, Facebook posts, conferences — everywhere.

● If a colleague from an underrepresented group can speak on an issue you’ve been asked to speak about, whether at a conference or for a quote in an article, give them the opportunity. If they decline, ask why — they may be interested but want PR training.

● Giving announcer and moderator positions to people from underrepresented groups at conferences is an easy way to start including them more. That’s not a license to not consider diversity and inclusion elsewhere else, but it is a step. Many people are very capable of being great emcees and moderators, so there’s little or no reason not to use this opportunity to include them. Note that EAG Boston and San Francisco 2017 both had a white man as the emcee.

● Don’t dismiss or trivialize the altruistic concerns ordinary people have. It’s great that people care about immigration reform, worms in children in impoverished regions, and dog rescue. It would be even more great if they put their energy into efforts of greater impact, but moving them in a more effective direction, whether within their currently preferred project or cause or to another, is easier done if they have a sense of community with you, which is easier achieved if they know you care about the issues they care about. It’s all but impossible to achieve if you stick your nose in the air at their altruism because their thinking on the weighty and new topic of effectiveness is underdeveloped — like yours once was.

● Quit the hero worship. Major progress is made by groups, not individuals. People should be praised for their individual contributions, and some people will be leaders, but that doesn’t mean other people aren’t contributing as much or more. Hero worship in EA is almost always directed towards white men, and while it’s great to celebrate their achievements, overdoing that celebration exacerbates the issue of how we represent ourselves to newcomers and outsiders, and encourages a masculine, individualistic culture where newcomers can’t thrive.

● Do not consider anyone’s arguments or positions above questioning or criticism. Never presume that someone has no place questioning someone else whose intellect you laud. No one is infallible, and no one has every answer or has considered every possible angle and argument. This particular form of hero worship is a common complaint from people who feel excluded from the community.

People from underrepresented groups: Own your worth. Don’t apologize for an A- job while others spin their C’s as A’s. Take credit for your work, even if you don’t personally want it, because other people like you need to see your success. Don’t do the dishes when that’s someone else’s responsibility this week. Apply for the jobs you want, not just the ones you are explicitly fully qualified for, because they’re written under the assumption that people are going to apply even when they only meet half the requirements — women don’t apply to jobs unless they meet all the posted requirements, whereas men apply when they meet 60%.

When possible, which is the vast majority of the time, use ordinary phrasing instead of jargon, at least with people who have only recently become involved.

● Stop interrupting people. Men are much more likely to interrupt than women are, and more likely still to interrupt women than other men. Not only does this disproportionately disempower women, but it’s rude and off-putting to everyone.

When people are interested in talking through something they’ve been thinking about in EA, have a conversation about it, even if you’ve already resolved your own thoughts on the topic and even if you don’t think there’s anything for you in the conversation. The other person will likely end up more informed and feel more welcomed, and it won’t take too much of your time. Remember too that being willing to engage with newcomers and people of lower status or perceivable “usefulness” is very common in other communities, and particularly advocacy communities, so when people act otherwise it seems surprising, rude and alienating.

● Don’t emphasize earning to give too much. This has been an ongoing discussion, and I think we’re slowly doing better.

● Be just as welcoming with people who do direct work on non-priority problems as you are with people who work in finance or tech. Not only can people contribute a lot more to the community and movement than their income, but keep in mind too that finance and tech specifically are places with particularly bad reputations for their exclusion of women and other historically marginalized people.

● Emphasize that doing the most good will necessarily mean different things for different people. Even if we ourselves know we’re speaking in generalities, it can very easily come off like we’re advocating a one-size-fits-all approach, or asserting that any one cause or career path is the best path to maximum impact for everyone.

● Represent the community’s values accurately. This can be a challenge in a single 140 character tweet, but not in a whole Twitter feed or in a conversation. Consistently presenting anti-malarial nets as the community’s primary concern is going to attract people with that particular interest. This means a relatively high proportion of people who are resistant to pushing new frontiers, as global poverty is a popular cosmopolitan cause that normal people can get lots of praise for contributing to more effectively, with no or nowhere near the personal risk of less mainstream causes and projects. Such an emphasis will also mean the community gets skipped over by people who are exploring other and unconventional ways of doing good.

Relatedly, our public image can and should be weird in the right ways. It can say true, abstract, challenging things like “We should consider the interests of all sentient beings,” “We don’t have all the answers, our goal is to find and implement them,” “How our actions affect people in the far future could vastly outweigh the impact they have now,” and “New technologies may transform the quality of life on Earth and beyond to a much greater extent than they have even in the past century.” And it can do all that without using jargon, without throwing around the term “AI” with no qualification or explanation, without looking or sounding like a young socially awkward white guy in tech, and while emphasizing the altruism motivating these intellectual explorations and providing palatable examples of relatively high-impact actions people can take — including, but not inordinately emphasizing, those that best help individuals in poverty. It’s not a question of either being weird AI fanboys or mainstream philanthropists.

● Don’t get hostile in conversations. Keep the focus on the information and arguments at hand.

● Don’t reward people for aggressive communication styles. If you want to express agreement with their content, but their delivery is bad form, you can say for instance “I agree, but your [snarkiness, ad hominem comment, exaggeration, etc] was unnecessary and not conducive to rigorous discussion.”

● Do not disproportionately penalize women for aggressive communication styles. When a man and woman are equally aggressive, people tend to see the man as more persuasive but the woman as less credible, and women are given feedback that they’re “too aggressive” three times as often as men. Both positions seem highly unlikely to line up with reality and are more likely unconscious efforts to punish nonconformity to gender stereotypes.

Relatedly, if you find yourself judging that a woman is too emotional, consider the men you know who are confrontational, who argue aggressively, who have expressed strong feelings about people they don’t know well, who can’t work well with attractive women, who jump to conclusions based on unexamined intuitions, who are obsessed with obtaining status, who are snarky, who level insults at others regularly, or who stoop to pissing contests. If you’re in the EA community, you know lots of men who demonstrate multiple such tendencies. In all likelihood men just hide their emotions better than women, which does not mean their judgements are less emotionally-motivated. It’s even possible that men’s judgements are more emotionally-motivated, as girls and women in society tend to have more social encouragement and opportunity to examine their emotions.

● Replace competitiveness with collaborativeness. In successful communities, people empower each other and become better off on the whole for it — another EA’s success strengthens and grows the community, and the community’s strength and size helps you and your purposes. So: Is someone’s counter to your argument making you feel defensive? This is an opportunity to get closer to the truth, together. Is someone considering starting a project that you were also thinking of? Combine your resources, and if it needs just one leader, sort out who’s best positioned for it — that’s great for the project. Are your donors shifting funds to a new organization? Sounds like you should drop inferior programs, and also like the community needs to grow the donor pool.

● Don’t try to take shortcuts to status, and particularly don’t try to gain status by disempowering other people. Status for most of us is not a zero-sum game. In fact, there is a lot of status to be gained by developing a reputation as someone who empowers other people. So it doesn’t matter what you’ve accomplished, you are not above giving a few minutes to an enthusiastic new EA who wants to learn how to get more involved, or at the least directing them warmly to someone who has more time to engage. And your public/​semi-private conversation at an EA event is not so important that you can’t take a few seconds to say hello to someone trying to enter the conversation and fill them in, or to change the topic for the new person — you can pick up the other conversation again later.

Relatedly, empower people, don’t use them — act in good faith, and show faith in your community members. Consider the other people in the community your collaborators, neither your competition nor a means to your ends. When collaborating with other EAs, be honest about your information, goals, and thought processes. Even if, for instance, you really just want someone to donate to or work for your project, and they’re deciding between yours and another, you should still give them your honest thoughts — or a better source — and critical or full information on the tradeoffs you see, not just what you think will convince them to support you instead of the other project. Introduce them to people at the other project if they aren’t introduced already. Help them make their own decision. Doing otherwise incentivizes further dishonesty and manipulativeness in the community.

Relatedly, consider the bigger picture, in everything you do. The good you can do does not just encompass the direct impact of your actions, but also how they influence other people. Establishing stronger norms of honesty would both incentivize stronger norms of intellectual rigor and select more strongly for new members who are intellectually rigorous rather than manipulative or manipulable. It’s also helpful to probably everyone as individuals to have a variety of people out there who appreciate you and will be enthusiastic about lending you a hand when you choose to ask for one, so be careful handling fire around bridges.

Similarly, when considering whether to go vegetarian or take some other step to avoid participating in a major moral problem, consider how not doing so could validate and perpetuate the biases and selfishness that enable people commit that act normally, and how that act could help others feel licensed to do other selfish and harmful things that you disagree with, like lying to sexual partners about having an STI or being dishonest and uncharitable in representations of your organization or preferred cause area.

Some people may have their own reasons for thinking that it’s good for them to act in and use people in short-sighted ways, and to be confident that they have nothing left to learn and no need to build social capital, but even if that actually is the right call for them individually, such short-sighted self-interest is bad for the broader EA community and limits what it can accomplish, so it should be discouraged. Controversy here may point to a deeper issue, of which I have seen concerning evidence, of some people using the broader EA community as a mere conduit to their preferred issue rather than a meeting place for everyone to learn from each other and help each other and grow the broader community and each other’s sub-communities on the whole. The community has a lot of room to grow, and actively trying to cannibalize each other is probably not in anyone’s long-run interest. So when, for instance, newcomers ask me about AI safety, I give them a clear and palatable introduction and I answer their questions or direct them to people who can answer better, and I do so even if we might not get a chance to talk about things I suspect would be a better use of their resources and which I have resolved are better use of mine. For me, the EA community isn’t just another place to pitch animal advocacy, it’s a place where I can learn and grow as an effective altruist, and where I can help others learn and grow as effective altruists. It’s a place where, in its better moments, people do good together, not alone.

● Give to the people in your community. Acknowledge their contributions, introduce them to people they might be interested in knowing, offer them your expertise, help them when they need a favor… this community is no exception to all communities’ needs for basic positive social norms.

● When people make mistakes, kindly and clearly identify them. If the mistake was not just an intellectual error but harmed someone, identify it in the interest of achieving justice for the person who was wronged, but also and perhaps more importantly in the interest of helping the person who made the mistake to grow and improve. That is to their benefit, the benefit of the community, the benefit of other people they would have gone on to wrong, and the benefit of others still who they’d be failing to help by falling short of who they could be. Encourage and reward good behavior privately and publicly, and discourage bad behavior privately, and more publicly and severely as it becomes more necessary to raise the costs to people of refusing to adopt better attitudes and behaviors. If we are only concerned with the direct impact of our own actions or don’t care about our omissions, we won’t get far in improving this community — we need to empower others to do better as well, so give people a genuine chance to improve.

To be clear, I’m not suggesting endless second chances, and some actions taken even once will warrant zero tolerance and immediate expulsion.

Also, even people who are exceptionally humble and exceptionally interested in personal growth still need to feel accepted and their egos can still be wounded, so take care not to overload people — give criticisms seriously but compassionately, focus on priorities, be clear about what happened, why the action was a problem, and what you think the person should have done instead, and in normal circumstances it’s probably best to give criticisms sparingly. Criticisms also have more credibility and are less hurtful when the critic has gained the respect and camaraderie of the criticized.

● Accept that you will make mistakes, and take responsibility when you do. Encourage yourself to value humility and growth even if it hurts your pride. We all make mistakes! When we are informed or otherwise realize that we have, we should take responsibility, rather than ignore the mistake or defend it and lose the opportunity to improve — not to mention incentivizing others to prioritize their own pride over self-improvement. Especially if you can feel an accusation wounding your ego and alerting your defenses, or if you can’t explicitly argue against an accuser’s points, you are probably not thinking very clearly. It should be a norm in the community to comfortably and casually admit “oh, you’re right I got that wrong” and “good point, I’ve changed my mind” and “I was not thinking about that effect of my actions, I’m sorry and thank you for bringing this to my attention.”

● Take up that humility more generally. Don’t judge that you’re right and another party is wrong before ensuring you know their reasoning — ask someone why they hold the position they do, maybe they’ve thought of something you haven’t just as you may be assuming you’ve thought of things they haven’t.

● You can disagree with people while entirely respecting their positions, appreciating their contributions, and recognizing them as an ally. The reason I spend my time strategizing to bring down animal farming and to expand humanity’s moral circle instead of working — directly at least — on AI safety generally seems to come down to intuitive differences between myself and people who prioritize direct work on AI safety. These differences are sometimes minor and in my experience generally irreconcilable with available information. I also disagree that near-term interventions to help individuals in poverty are the best use of most EA’s resources because I don’t think lives matter as much as well-being, and poverty interventions are not relatively robust in their address of well-being. I disagree with many of my allies and colleagues about the value of farmed animal welfare reforms and other near-term interventions ultimately because I tend to be more risk-tolerant and compelled by expected value than they are, and because I consider the net impacts of near-term interventions sufficiently uncertain that I don’t think it’s useful to consider them categorically more measurable than interventions whose intended impacts are less direct or further in the future.

Nonetheless, I’m very excited that these people are working on these projects, which I still consider important even if I disagree that they’re the best use my or these individuals’ resources, and I still have a lot of respect for some of these allies’ and colleagues’ analyses, and I am deeply moved by their altruistic drives and grateful for their contributions to the EA community and to my own thinking on these issues. Disagreement is critical for finding the best answers to the kinds of questions EAs ask.

● There is a point at which championing “free speech” actually inhibits it, enabling what was once innovative, challenging, rigorous discussion to become regressive, harmful, thoughtless trolling and/​or identity politics. When people say severely intolerant things that disenfranchise other people — especially if they for instance cannot justify it, respond to criticisms of it with aggressive repetition of their claims with no evidence and/​or with personal attacks, and cannot explain why it’s important that they say it at all — don’t tolerate it.

For instance, it should be outright unacceptable for someone to say that women do not contribute to society and are leeches if they don’t offer men sex. This actually happened, recently, and is a problem for two reasons: One, the factual claim is highly contrary to economic and other data as well as extensive anecdotal evidence, and such unrigorousness should be discouraged. Two, the value judgement, which is explicitly sexist to an atypically extreme degree, is well beyond the limit of what the community should accept as any kind of a “diversity of opinion” unless we want to severely limit our diversity of participants, and as such that very diversity of opinion. People are both less able to and less interested in contributing their resources to the community when they are treated with such hostility and when such hostility is accepted by the community. It is women’s interests to assume that every man who is okay with this person’s behavior has an appallingly poor understanding of sexism in society, if not also of basic social norms generally, and that as such he probably harbors a dangerous level of sexism himself, if not also a shockingly — contextually — limited intellectual capabilities given the obvious lack of intellectual rigor in the offender’s comments. So toleration of such comments makes the whole community look highly unappealing.

Happily, this particular individual — who is probably a troll in general — was banned from the groups where he repeatedly and unrelentingly said such things, though it’s concerning there was any question about whether this was acceptable behavior. Maybe we should have a reference document of what kinds of actions in online forums warrant an explanation of the problem, ensuing non-engagement, warnings from the moderator, and bans.

To be clear, by tolerating rude and intellectually unrigorous behavior we are in fact choosing to have such people in the community in the place of the more rigorous and compassionate people they are likely to put off. Such toleration of intolerance is also likely to normalize that intolerance and as such to increase the biases in the rest of community. It concerns me that I even have to bring this up as a problem, as I think e.g. most Fortune 500 companies have by now figured out that it’s very important that employees not be outright assholes to other employees. [Edit: example that came to mind redacted because while problematic, I would not describe the person as an “outright asshole,” though this action was still a serious problem.] Yes, some people in broader society now respond to correctable offenses with a mob mentality and too much readiness for ostracization, but just because some people have swung too far past the mark doesn’t mean we should default to a status quo that falls so short of it.

Hiring processes and employee management are a big topic, but for starters, take care with job postings: Use less masculine language; talk about the concrete skills and experience you’re interested in instead of appealing to people with “startup” experience; ensure that the qualities you say are “required” are actually required; and appreciate that women may conceive of their achievements differently than men tend to, for instance attributing their successes more to their team rather than to themselves.

● Men, accept that many women will be your equals, and others your superiors, in intelligence, knowledge, and other abilities you aspire to or pride yourself in. Even those who aren’t will sometimes have a better argument or more relevant information than you. And no, just because you can point to one or two women whose intellects and other competencies you appreciate does not mean you are evaluating other womens’ fairly — especially if the women you are thinking of are in your community and share your positions. The same goes for people of color, and others.

● People who belong to currently disenfranchised groups, adopt the attitude that the success of other people who are disenfranchised, particularly for the same reasons as you, is your success. Women who encounter discrimination early in their careers may distance themselves from other women, refuse to help them, and align themselves with men at other women’s expense. The disempowerment of women in the EA community may make women feel as though there is only room for a few women to have some voice, but we don’t need to accept someone else’s narrative that we have to compete with each other — we can make more room for each other, like women in other masculine men-dominated communities have done before us and are doing alongside us, by empowering each other. As I’ve said already, this is not a zero-sum game: Every person of color’s success should, with sustained inclusionary efforts from the rest of the community, reduce some racism in the community, which in turn increases opportunity for other people of color in a virtuous circle.

● Mentor people from underrepresented groups. Or if you belong to an underrepresented group, seek out mentors.

● Take an interest in people. You will at times, often even, have to judge when someone isn’t going to be so involved in the movement that it’s worth your time to continue engaging, but give people a chance, and try to be mindful of your intuitions, some of which will be more valid and useful than others and some of which will be plain biased — try to be conscientious in that judgement and focus on concrete measures of a person’s likelihood to engage well enough that they’ll learn to do good better.

Finally: Take responsibility for improving diversity and inclusion in EA. Whatever your role in the community and movement and however inclusive your actions tend to be already, there is more you can do, and saying it’s someone else’s problem to solve will only result in a collective action problem..

Other notes

See also Kelsey Piper’s notes on failure modes in efforts to increase demographic diversity, Julia Wise’s post on specific actions people can take to be more welcoming at events, and Owen Cotton-Barrett’s post on being welcoming.

I should note that I put vastly more time and effort into working with people outside of EA to develop their thinking on effectiveness independently of the EA community than I do bringing new people into the community, which frankly I only do when they’ve explicitly expressed interest. This is because I usually expect introducing them to the community to waste their time, cause them stress, cost some of my relationship with them because of that, and most importantly, turn them off from thinking about effectiveness. In fact, I think we backfire often just because we present ourselves so suboptimally.

The time I have spent on EA community-building, which has been substantial, has focused on supporting individuals who are already in the community, for the most part in the wing that intersects with the animal advocacy community. I should note, brusque though this comment may be, that the animal-advocacy-focused sub-community of EA tends to be significantly more socially competent, welcoming, and proficient in the kinds of inclusionary practices I’ve suggested here than some other parts of the community. This may be largely explained by how women-dominated the animal advocacy community is — though heavily white and guilty of other failings — and how its members are generally much better versed in issues of discrimination and inclusion than EAs are. Animal advocates, particularly in the farmed animal wing, tend to be highly liberal and generally actively encourage concern for broad social justice — which stands in stark contrast to the many people in the EA community who use strawmans and the worst of the social justice community to dismiss, insult, and otherwise actively discourage any association with the term, to the point of taking pride in that opposition.

I should also note that most other women, white and of color, who have been in the community for several years and who I have spoken with about diversity and inclusion issues, are exhausted from talking about and even thinking about this problem for so long and to so little avail. True, the vast majority of that conversation has been in private or otherwise sequestered discussions, and mostly among people who agree there’s a problem and aren’t contributing to it as much as others, whether by act or omission. That’s why I’m putting all of these thoughts online. Regardless, people from more represented backgrounds and who are otherwise in more influential positions need to take up this mantle.

Also FYI, I am currently reading and taking notes on What Works: Gender Equality by Design and intend to share its insights — even if they’re potentially somewhat cherry-picked and otherwise weaker evidence than we’d like, as pop science books often are — hopefully within the next month or so.

Thank you Jennifer Fearing for the handful of suggestions I took from your advice to animal advocates on how to promote gender inclusion in animal advocacy leadership.