The minimization and denial among these comments is horrifying.
I am a female AI researcher. I come onto this forum for Neel Nanda’s interpretability research which has recently been fire. I’ve experienced abuse in these communities which makes the reaction here all the more painful.
It is appropriate to minimize things which are in fact minimal. The majority of these issues have been litigated (metaphorically) before. The fact that they are being brought up over and over again in media articles does not ipso facto mean that the incident has not been adequately dealt with. You can make the argument that these incidents are part of a larger culture problem, but you have to actually make the argument. We’re all Bayesians here, so look at the base rates.
The one piece of new information which seems potentially important is the part where Sonia Joseph says, “he followed her home and insisted on staying over.” I would like to see that incident looked into a bit more.
Given the gender ratio in EA and rationality, it would be surprising if women in EA/rationality didn’t experience more harassment than women in other social settings with more even gender ratios.
Consider a simplified case: suppose 1% of guys harass women and EA/rationality events are 10% women. Then in a group of 1000 EAs/rationalists there would be 9 harassers targeting 100 women. But if the gender ratio was even, then there would be 5 harassers targeting 500 women. So the probability of each woman being targeted by a harasser is lower in a group with more even gender ratio. For it to be the case that women in EA/rationality experience the same amount of harassment as women in other social settings the men in EA/rationality would need to be less likely to harass women than the average man in other social settings.
It is also possible that the average man in EA/rationality is more likely to harass women than the average man in other social settings. I can think of some reasons for this (being socially clumsy, open to breaking social norms etc) and some against (being too shy to make advances, aspiring to high moral standards in EA etc).
While many of these claims are “old news” to those communities, many of these claims are fresh. The baseline rate reasoning is flawed because a) sexual assault remains the most underreported crime, so there is likely instead an iceberg effect, and b) women who were harassed/assaulted have left the movement which changes your distribution, and c) women who would enter your movement otherwise now stay away due to whisper networks and bad vibes.
While many of these claims are “old news” to those communities, many of these claims are fresh.
Can you clarify which specific claims are new? A claim which hasn’t been previously reported in a mainstream news article might still be known to people who have been following community meta-drama.
The baseline rate reasoning is flawed because a) sexual assault remains the most underreported crime, so there is likely instead an iceberg effect,
I’m not sure how this refutes the base rate argument. The iceberg effect exists for both the rationalist community and for every other community you might compare it to (including the ones used to compute the base rates). These should cancel out unless you have reason to believe the iceberg effect is larger for the rationalist community than for others. (For all you know, the iceberg effect might be lower than baseline due to norms about speaking clearly and stating one’s mind.)
b) women who were harassed/assaulted have left the movement which changes your distribution,
Maybe? This seems more plausible to confound the data than a) or c), but again there are reasons to suppose the effect might lean the other way. (Women might be more willing to tolerate bad behavior if they think it’s important to work on alignment than they would tolerate at say, their local Magic the Gathering group).
c) women who would enter your movement otherwise now stay away due to whisper networks and bad vibes.
Even if true, I don’t see how that would be relevant here? Women who enter the movement, get harassed, and then leave would make the harassment rate seem lower because their incidents don’t get counted. Women who never entered the movement in the first place wouldn’t affect the rate at all.
Strong upvote. As another female ai researcher: yeah, it’s bad here, as it is everywhere to some degree.
To other commenters, especially ones hesitant to agree that there have been problems due to structural issues, claiming otherwise doesn’t make this situation look better—the local network of human connections can only look good to the outer world of humans by being precise about what problems have occurred and what actual knowledge and mechanisms can prevent them. you’re not gonna retain your looking good points to the public by groveling about it, nor by claiming there’s no issue; you’ll retain looking good points by actually considering the problem each time it comes up, discussing the previous discussions, etc. (though, of course, like, efficiently, according to taste. Not everyone has to write huge braindumps like I find myself often wanting to.) nobody can tell you how to be a good person; just be one. put your zines in the slot in the door and we’ll copy em and print em out. but dont worry about making a fuss apologizing; make a fuss explaining a verifiable understanding.
Some local fragments of social network are somewhat well protected by local emotional habits; but many subgroups have unhealthy patterns even now, years and years after the event in the article.
So, to minimizing commenters, don’t dismiss the messenger; people being upset are not a reason for you to be upset too, it’s a reason for you to have sympathy and take time to think. You’re safe from this criticism with me, and I will argue for you to be safe from it for the most part even if you’ve caused harm, because I believe in reparative and restorative justice; victims aren’t safe until there’s a network of solidarity that prevents the harm in the first place and the process of getting there can be accelerated by frequent sources of harm getting to [pretend to not be causing harm]{comment 7mo later: ehh, I mean more like, others won’t apply current typical retribution; not hiding of the harm} if they’ll help as hard as they can with preventing anyone from ever harming another again—starting with themselves.
(Though it’s understandable if you’re sympathetically upset or worried about being a victim. These things are less gendered than popular portrayal implies.)
In the local words: Simulacrum 3 is not optional, you just gotta ensure you can promise to be on Simulacrum 1 when discussing issues and your views about how to deal with them.
To whistleblower67, some commentary on why I’m here: very few communities live like the future we should have, though some institutions are better at preventing abuse than others—I imagine there are institutions the reader thinks of as good, maybe better than the lesswrong crowd or maybe not, and I would caution that many of those institutions also have abuse patterns. eg, universities come to mind as probably having a slightly lower rate of abusers but still very much not zero, except, now that I say that I’m not even confident. I initially hoped that it would be less terrible here in the rat community; it hasn’t been. it’s the same problem as every other online network I’ve been in, and this one has it about as bad as a typical online network of emotional connections, maybe worse sometimes, maybe better others. Which is to say, fuckin awful, not to be minimized, we don’t heal all communities magically with a snap of our fingers. I want to feel comfy here—and I don’t.
And also, I’m not leaving the research field just because nobody agentically abusive wants to fess up to having been an agentic abuser. If human-scale community abusers are enough to stop our research on inter-agent coprotective alignment, that seems bad, especially if some of the people working on inter-agent coprotective alignment don’t care to actually work on it because they themselves are misaligned with decency! It seems to me that these situations are issues of the interpersonal alignment problem not being solved: how do we design a system where the members of that system can learn to trust each other when needed and distrust when it isn’t valid, so that if someone is abused, the victim can speak out against the abuser safely, and yet the abuser cannot use this same system against the victim.
If you have suggestions for other forums I should also visit to discuss the mechanical details of coprotective ai design as an independent researcher, I’d love to hear them, ideally in DM so the annoying boys don’t join too quick. but there are plenty of ladies roughing it here in the hope of making a difference, plenty of “people of gender” so to speak, in general. We can protect each other, but doing so is not trivial; don’t join this community blindly, but don’t join any community blindly. To build solidarity, one must be ready to defend those in need.
and a quick demo of what this looks like when I don’t feel so agitated by the topic:
The links from my referenced comment, pasted here for others passing by:
https://www.microsolidarity.cc/articles/cults—this site is in general my favorite site on the internet right now; it discusses how to create healthy co-supportive, solidarity-heavy social groups without accepting domination or cutting off friendships. Domination and isolation being a key component of cults.
it is the system of unfairness that must end, not the people in it, even on top. no death penalty, no penalty of forced labor, no penalty of total confinement, for causing harm to others, in my view. Temporary confinement and fair representation of harms to the social network instead. We need to change how these systems of justice work fundamentally in order to make them durable to the degree needed to protect against the coming era.
It seems from your link like CFAR has taken responsibility, taken corrective action, and states how they’ll do everything in their power to avoid a similar abuse incident in the future.
I think in general the way to deal with abuse situations within an organization is to identify which authority should be taking appropriate disciplinary action regarding the abuser’s role and privileges. A failure to act there, like CFAR’s admitted process failure that they later corrected, would be concerning if we thought it was still happening.
If every abuse is being properly disciplined by the relevant organization, and the rate of abuse isn’t high compared to the base rate in the non-rationalist population, then the current situation isn’t a crisis—even if some instances of abuse unfortunately involve the perpetrator referencing rationality or EA concepts.
The minimization and denial among these comments is horrifying.
I am a female AI researcher. I come onto this forum for Neel Nanda’s interpretability research which has recently been fire. I’ve experienced abuse in these communities which makes the reaction here all the more painful.
I don’t want to come onto this forum anymore.
This is how women get driven out of AI.
It is appropriate to minimize things which are in fact minimal. The majority of these issues have been litigated (metaphorically) before. The fact that they are being brought up over and over again in media articles does not ipso facto mean that the incident has not been adequately dealt with. You can make the argument that these incidents are part of a larger culture problem, but you have to actually make the argument. We’re all Bayesians here, so look at the base rates.
The one piece of new information which seems potentially important is the part where Sonia Joseph says, “he followed her home and insisted on staying over.” I would like to see that incident looked into a bit more.
Given the gender ratio in EA and rationality, it would be surprising if women in EA/rationality didn’t experience more harassment than women in other social settings with more even gender ratios. Consider a simplified case: suppose 1% of guys harass women and EA/rationality events are 10% women. Then in a group of 1000 EAs/rationalists there would be 9 harassers targeting 100 women. But if the gender ratio was even, then there would be 5 harassers targeting 500 women. So the probability of each woman being targeted by a harasser is lower in a group with more even gender ratio. For it to be the case that women in EA/rationality experience the same amount of harassment as women in other social settings the men in EA/rationality would need to be less likely to harass women than the average man in other social settings. It is also possible that the average man in EA/rationality is more likely to harass women than the average man in other social settings. I can think of some reasons for this (being socially clumsy, open to breaking social norms etc) and some against (being too shy to make advances, aspiring to high moral standards in EA etc).
While many of these claims are “old news” to those communities, many of these claims are fresh. The baseline rate reasoning is flawed because a) sexual assault remains the most underreported crime, so there is likely instead an iceberg effect, and b) women who were harassed/assaulted have left the movement which changes your distribution, and c) women who would enter your movement otherwise now stay away due to whisper networks and bad vibes.
Can you clarify which specific claims are new? A claim which hasn’t been previously reported in a mainstream news article might still be known to people who have been following community meta-drama.
I’m not sure how this refutes the base rate argument. The iceberg effect exists for both the rationalist community and for every other community you might compare it to (including the ones used to compute the base rates). These should cancel out unless you have reason to believe the iceberg effect is larger for the rationalist community than for others. (For all you know, the iceberg effect might be lower than baseline due to norms about speaking clearly and stating one’s mind.)
Maybe? This seems more plausible to confound the data than a) or c), but again there are reasons to suppose the effect might lean the other way. (Women might be more willing to tolerate bad behavior if they think it’s important to work on alignment than they would tolerate at say, their local Magic the Gathering group).
Even if true, I don’t see how that would be relevant here? Women who enter the movement, get harassed, and then leave would make the harassment rate seem lower because their incidents don’t get counted. Women who never entered the movement in the first place wouldn’t affect the rate at all.
Strong upvote. As another female ai researcher: yeah, it’s bad here, as it is everywhere to some degree.
To other commenters, especially ones hesitant to agree that there have been problems due to structural issues, claiming otherwise doesn’t make this situation look better—the local network of human connections can only look good to the outer world of humans by being precise about what problems have occurred and what actual knowledge and mechanisms can prevent them. you’re not gonna retain your looking good points to the public by groveling about it, nor by claiming there’s no issue; you’ll retain looking good points by actually considering the problem each time it comes up, discussing the previous discussions, etc. (though, of course, like, efficiently, according to taste. Not everyone has to write huge braindumps like I find myself often wanting to.) nobody can tell you how to be a good person; just be one. put your zines in the slot in the door and we’ll copy em and print em out. but dont worry about making a fuss apologizing; make a fuss explaining a verifiable understanding.
Some local fragments of social network are somewhat well protected by local emotional habits; but many subgroups have unhealthy patterns even now, years and years after the event in the article.
So, to minimizing commenters, don’t dismiss the messenger; people being upset are not a reason for you to be upset too, it’s a reason for you to have sympathy and take time to think. You’re safe from this criticism with me, and I will argue for you to be safe from it for the most part even if you’ve caused harm, because I believe in reparative and restorative justice; victims aren’t safe until there’s a network of solidarity that prevents the harm in the first place and the process of getting there can be accelerated by frequent sources of harm getting to [pretend to not be causing harm]{comment 7mo later: ehh, I mean more like, others won’t apply current typical retribution; not hiding of the harm} if they’ll help as hard as they can with preventing anyone from ever harming another again—starting with themselves.
(Though it’s understandable if you’re sympathetically upset or worried about being a victim. These things are less gendered than popular portrayal implies.)
In the local words: Simulacrum 3 is not optional, you just gotta ensure you can promise to be on Simulacrum 1 when discussing issues and your views about how to deal with them.
To whistleblower67, some commentary on why I’m here: very few communities live like the future we should have, though some institutions are better at preventing abuse than others—I imagine there are institutions the reader thinks of as good, maybe better than the lesswrong crowd or maybe not, and I would caution that many of those institutions also have abuse patterns. eg, universities come to mind as probably having a slightly lower rate of abusers but still very much not zero, except, now that I say that I’m not even confident. I initially hoped that it would be less terrible here in the rat community; it hasn’t been. it’s the same problem as every other online network I’ve been in, and this one has it about as bad as a typical online network of emotional connections, maybe worse sometimes, maybe better others. Which is to say, fuckin awful, not to be minimized, we don’t heal all communities magically with a snap of our fingers. I want to feel comfy here—and I don’t.
And also, I’m not leaving the research field just because nobody agentically abusive wants to fess up to having been an agentic abuser. If human-scale community abusers are enough to stop our research on inter-agent coprotective alignment, that seems bad, especially if some of the people working on inter-agent coprotective alignment don’t care to actually work on it because they themselves are misaligned with decency! It seems to me that these situations are issues of the interpersonal alignment problem not being solved: how do we design a system where the members of that system can learn to trust each other when needed and distrust when it isn’t valid, so that if someone is abused, the victim can speak out against the abuser safely, and yet the abuser cannot use this same system against the victim.
If you have suggestions for other forums I should also visit to discuss the mechanical details of coprotective ai design as an independent researcher, I’d love to hear them, ideally in DM so the annoying boys don’t join too quick. but there are plenty of ladies roughing it here in the hope of making a difference, plenty of “people of gender” so to speak, in general. We can protect each other, but doing so is not trivial; don’t join this community blindly, but don’t join any community blindly. To build solidarity, one must be ready to defend those in need.
and a quick demo of what this looks like when I don’t feel so agitated by the topic:
Cults are bad actually, and I often say so with links to cult resistance education on this forum when people fret about cults or even giggle about cults being fun, because of the issues that have occurred around here from people thinking trusting authorities is a good idea.
The links from my referenced comment, pasted here for others passing by:
it is the system of unfairness that must end, not the people in it, even on top. no death penalty, no penalty of forced labor, no penalty of total confinement, for causing harm to others, in my view. Temporary confinement and fair representation of harms to the social network instead. We need to change how these systems of justice work fundamentally in order to make them durable to the degree needed to protect against the coming era.
It seems from your link like CFAR has taken responsibility, taken corrective action, and states how they’ll do everything in their power to avoid a similar abuse incident in the future.
I think in general the way to deal with abuse situations within an organization is to identify which authority should be taking appropriate disciplinary action regarding the abuser’s role and privileges. A failure to act there, like CFAR’s admitted process failure that they later corrected, would be concerning if we thought it was still happening.
If every abuse is being properly disciplined by the relevant organization, and the rate of abuse isn’t high compared to the base rate in the non-rationalist population, then the current situation isn’t a crisis—even if some instances of abuse unfortunately involve the perpetrator referencing rationality or EA concepts.
Sorry you experienced abuse. I hope you will contact the CEA Community Health Team and make a report: https://forum.effectivealtruism.org/posts/hYh6jKBsKXH8mWwtc/contact-people-for-the-ea-community