After the Time article on sexual harassment and abuse within Effective Altruism and Owen Cotton-Barratt's resignation there's been a lot of discussion around the extent to which EA culture did or didn't contribute. This has included proposals that EA try to be more conventional, such as discouraging polyamory or hookups within the community. At which point others naturally bristle: what happens between consenting adults is no one else's business, and it's wrong to try and influence others here. For example:

I don't see any way you could meaningfully "address" the work/social overlap without trying to get people not to date, live with or befriend people they otherwise would have dated, lived with, or befriended. And if you put it in those terms, it seems messed up, right?


I will not be a part of community which treats conscious and consensual behavior of adult people as their business.


Perhaps consider minding your own business about whether or not consenting adults sleep together.

Now, the people who wrote these have more nuanced views than these quotes suggest, and I think we don't actually disagree much when it comes to views on detailed situations. Still, I think it's worth getting some of that nuance out there.

Specifically, consent isn't always enough. Even consent plus a mature awareness of power dynamics isn't always enough. Consider a case of someone whose job it is to direct funding for a foundation sleeping with one of the people who runs an organization they might recommend funding. And let's further imagine that the grantmaker and grantee both have a sophisticated understanding of power dynamics, great communication, solid introspection, strong self-confidence, and the best of intentions. Even then, this has corrosive effects on the community, including:

  • Other grantees would then feel pressure to sleep with grantmakers, leading to bad interactions, including ones where all the signals that the grantmaker receives are that the grantee wants this.

  • The funders behind the grantmaker may reasonably worry that the grantmaker's judgement is clouded by their otherwise positive views towards this grantee or that there was quid pro quo.

  • People with unsavory intentions may choose to become grantmakers because a norm of "it's ok to sleep with grantees" is very vulnerable to abuse.

These are harmful enough that we can't leave it to the couple to determine whether their roles as grantmaker/grantee and lovers are compatible. That sometimes someone's role conflicts with a relationship they would like to have is unfortunate, but not always avoidable and not unique to this case: other examples of romantic relationships with similar issues include ones between managers and their reports or professors and their students.

Organizations handle these conflicts in a range of ways, often depending on how they think about the effects of these relationships on others and whether they can satisfactorily mitigate the professional consequences of a particular relationship. For example:

  • A university might consider the initiation of any relationships between students and professors to have a harmful effect on their culture, and that students should be able to trust that professors are interested in them only academically. If they prohibit them, a professor would need to choose between avoiding such relationships and resigning their post.

  • A large funder might require disclosure and recusal: when a relationship relevant to a grantmaker's work comes up they work with their manager to transfer responsibilities so that the grantmaker is no longer influencing outcomes for the grantee. This is also how larger organizations usually handle employee relationships that interact with formal power.

  • A small funder might not be able to do this, perhaps because they only have one grantmaker in an area. Options for handling this could include committing not to fund the relevant organization, the grantmaker leaving, or the grantmaker passing up the relationship. All of these have significant downsides. Smaller organizations, or specialized departments within larger organizations, can have similar issues that would keep them from being able to mitigate the effects of an employee relationship.

There are also other areas where we accept people have a valid interest in discouraging some kinds of consensual interpersonal behavior, where power dynamics and conflicts of interest are not the concerns. For example, norms against cheating in relationships, sleeping with your bandmates, or hookups within a military unit.

As I wrote above, I don't think this is especially controversial. But it shows that the question is what norms we should have and not whether it is legitimate to have norms beyond consent. This is the approach we need when considering the actually controversial cases.

On the more specific question of what norms to have, I don't know. This requires weighing harms like the departure of people frustrated by excessive romantic attention, abuse of ambiguous power, loving relationships that don't happen, exclusion of unconventional thinkers, and delay in ejecting jerks. I think the discussion has shown both that we didn't have much agreement on what norms we'd been operating under and that we have a wide range of opinion on what's reasonable behavior. I'm hoping we can get closer to figuring out what sort of norms would best make a thriving community that supports all of us in making the world better.

Disclosure: my wife is on the Community Health Team at the Centre for Effective Altruism. I haven't run this post by her and don't know her views.

New Comment
16 comments, sorted by Click to highlight new comments since:

Relationships between grant-maker and grantee or professor and student are violations of professional norms. We rightfully blame the grant-maker and professor for them and we don't blame the grantee and student.

"Consent Isn't Always Enough" is a misleading phrasing to make this point: It mixes the personal and professional level. We may want a norm on the professional level that certain relationships are not accepted. The norm that there should be consent in the relationship happens on the personal level – we don't expect a manager to investigate consent in the relationships of their staff, nor is lack of consent mostly an internal disciplinary matter.


This isn't limited to professional contexts, though I do think some of the situations there are clearer.

An example norm for a context in which no organization is involved is that I don't think people (especially organizers) should be hitting on first-timers at EA meetups.

Not hitting on people on their first meetup is good practice, but none of the arguments in OP seem to support such a norm.

Perhaps less charitably than @Huluk, I find the consent framing almost tendentious. It's quite easy to see how the dynamics denounced have little to do with consent; here are two substitutions which show how the examples are professional ethics matters, and orthogonal to the intimacy axis:

- one could easily swap "sexual relations" with "access to their potential grantee's timeshare" without changing much in terms of moral calculus;
- one could make the grantee as the recipient of another, exclusive grant from other sources. In this case, flirting with a grantmaker would no longer have the downstream consequences OP warned about.

All in all, the scenario in OP seems to call not for more restrictive sexual norms, but for explicit and consistently enforced anti-collusion/corruption regulations.

Once again: this is limited to the examples provided by @jefftk, and the arguments accompanying them. It's possible that consent isn't always enough in some contexts within EA, for reason separated from professional ethics - but I did not find support for such thesis in the thread.

It mixes the personal and professional level

Possibly reflective of a wider issue in EA/rationalist spaces where the two are often not very clearly delineated. In that sense EA is more like hobby/fandom communities than professional ones. 

In my opinion, the more money (or other resources, or power) flows through someone's hands, the less excuse that person has for saying "hey, this is just a hobby, we do not want the boring professional norms".

Hobby is when you do it in your free time, and if someone does not respect your boundaries, you can easily find a different hobby. If it is your only or major source of income, when denying "consent" might mean losing your income, it is de facto a job.

And if someone believes that the people in positions of power who get lots of "consent" from their underlings are evaluating this situation impartially... I may have a bridge to sell you.

This is an interesting point. And the pitfalls become obvious when put in that context. 

Here is a pattern to watch out for: 

Generally accepted virtues and rules are there to keep bounded-rational humans from breaking through the guardrails of Chesterton/Schelling fences, and breaking through one fence, such as "we use consequentialism to be effective and won't stoop to supporting charities that cure rare diseases in cute puppies!" becomes a Murder-Gandhi slippery slope very easily. 

"Look, we are enlightened now, we don't need no stinking guardrails!" is an absolutely classic pitfall. I wonder if the EA community educates its members on the dangers of this.

And let's further imagine that [...] both have a sophisticated understanding of power dynamics, great communication, solid introspection, strong self-confidence, and the best of intentions.

Somewhat of a separate point: People may tend to overestimate these factors. A friend brought up the argument that it seems to them that people sleeping with each other often makes things awkward or creates a mess of some sort. If true, that point by itself, independently of any further reasoning about whether anyone is morally at fault and independently of questions about bad incentives around power dynamics, is an (impact-driven) argument to not do it in your professional community.

Everything is about sex except sex.  Sex is about power.  (attribution unclear, likely not O. Wilde)

I'm not part of any of the EA communities for multiple reasons, but from the outside this looks like some mix of the standard cult/intense-culture problem of convincing yourself that your work is massively outsized in importance, and anything that makes the work tolerable is therefore acceptable.  Combined with a bit of public overreaction to flaunting one's weirdness making the bad parts far more salient.  There are certainly problems, but asking for lifestyle changes or more professional separation between work and living is a bit of an isolated demand for rigor.

This branch of EA also tends to be a bit more literal than most cults/mission-oriented-enterprises, so it resists the most common and effective technique to avoid outside judgement of things they think is OK: hypocrisy.   Most high-profile businesses include office romances and abuse of power relations.  But they're hidden and handled quietly when they reach the level of causing un-ignorable problems.   

commenting on the body, separate from the incident that prompted this. when i was in school:

  • the occasional professor would invite the class for drinks after a test.
  • a subset of students i TA’d would invite me for tea.
  • a subset of students would bake food and share it after exams.

no mention of relationships yet. but all these activities are exactly those avenues by which people learn about each other and by which they form bonds. the professors i bonded with were exactly those professors whose office hours i attended most. and vice versa for the students i bonded with attending more of my office hours.

student/teacher bonding means the student is more comfortable asking the teacher for help, means the teacher better understands how to frame things in a way the student will get. if you ran the study, you would surely find correlation between this and course scores. does that mean this style of bonding is unethical?

if you ran the study and found that informal socializing didn’t decrease any student’s learning outcome, but it did increase some outcomes non-uniformly, would that be unethical?

it seems to me that the vast majority of times where relationships cause power problems is when one of the members is in competition with another. the argument that the criteria for competition among employees in the workplace shouldn’t involve sex is largely an argument that people shouldn’t be coerced into participating in a competition they don’t want to be a part of. aiming for consensus w.r.t. which criteria you should apply such that all workers prefer to be in such competition is a somewhat limited prospect. maybe there’s some progress that can be made there, but i expect the bulk of progress will actually be in finding ways to not force everyone into the same competition. in the post-scarcity world where you don’t have to compete in the workplace to stay alive, this consent issue would largely disappear. until then, the best we can do is fragment the competitive pools: less hierarchical workplaces such that the effect of any one competition is radically reduced, and more employment choices so that those who want work and life to be separate can avoid entering into direct competition with those who want work and life to overlap.