Iâm just puzzled about the apparent double standard where the first people to make allegations enjoy privacy & anonymity (even if their allegations seem to be largely false or exaggerated), but the people theyâre accusing donât enjoy the same privilege.
I agree that the Forumâs rules and norms on privacy protection are confused. A few observations:
(1) Suppose a universe in which the first post on this topic had been from from Nonlinear, and had accused Alice and Chloe (by their real names) of a pattern of mendaciously spreading lies about Nonlinear. Would that post have been allowed to stay up? If yes, it is hard to come up with a principled reason why Alice and Chloe canât be named now.
If no, we would need to think about why this hypothetical post would have been disallowed. The best argument I came with would be that Alice and Chloe are, as far as I know, people with no real prominence/âinfluence/âpower (âPIPâ) within or without EA. Under this argument, there is a greater public interest in the actions of those with PIP, and accepting a role of PIP necessarily means sacrificing some of the privacy rights that non-PIPs get.
(2) Another possibility involves the idea of standing. Under this theory, Alice and Chloe had standing to name people at Nonlinear because they were the ones who allegedly experienced harm. Ben had derivative standing because Alice and Chloe had given him permission to share their stories on the Forum. Under this theory, Nonlinear (and individuals named in Benâs post) would have standing to name Alice and Chloe as the parties allegedly harmed by their conduct. The rest of us would not have standing. [Edit to add footnote here.[1]]
Maybe the aggrieved party isnât the best imaginable arbiter of whether someone should be unmasked, but itâs probably better than concluding that each and every one of us get to make that decision. That could more easily lead to a situation in which 99% of us agree that the names shouldnât be disclosed, and the 1% gets to decide. Of course, that could happen under a standing theory as well, but the 1% minority opinion has to coincide with the list of parties with standing. That at least narrows the problem.
(3) This situation involves a mix of Forum and non-Forum conduct. I sense that the norm for not doxxing people is stronger for Forum conduct (or at least for conduct involving similar sorts of speech) than from non-Forum conduct. This may reflect a difference in the norms of Internet communities vs. physically-grounded communities, or a sense that Forum speech is often core expressive speech or otherwise entitled to extra solicitude.
In any event, based on the fact that a number of people seem to know their identities, it sounds like the tale of Alice and Chloe was not a closely-kept secret before Benâs post went up. Presumably this information came, directly or indirectly, from Alice and Chloe. In other words, there has been potentially relevant non-Forum conduct that did not magically morph into Forum conduct merely because Ben decided to write a post sharing the same information.
That raises the following hypothetical. Letâs say that Dan allegedly made some homophobic, racist, sexist, and/âor otherwise offensive comments.[2] Does the location in which the comments were made change whether he can be identified. For example, is it OK to write a Forum post identifying Dan by name if those comments were posted on the Forum under a pseudonym? What if they were said at an EAG afterparty? In a non-EA space on Reddit? With a mask on in public (by someone who recognized Danâs voice)? If the answers are not the same, why is that the case? I donât claim to know the answers, by the way; this isnât a trick question.
(4) I do not at present have a clear opinion on whether naming Alice, Chloe, or both should be allowed on the Forum. I do have an opinion that there needs to be a clear, logical set of rules or at least principlesâideally, laid out in advance as far as that is practical. I donât like the idea of apparently ad hoc decisions about who gets privacy protections and who does not.
[Added footnote: I should note that standing, which is used as a legal metaphor here, does not mean âis none of your business.â We often impose standing requirements to make sure the person making the decision has appropriate incentives, context, etc.
Here, the various people who A & C have allegedly spread lies about are in a much better position than I am to know the relevant facts. If they have not concluded that disclosure of A & Câs names is warranted, itâs not clear why Iâor any other reader of the relevant postsâwould do a better job as decisionmaker.
Another reason for limited standing is practicality. For instance, in US law, we donât usually allow suits where the alleged harm is âIâm a taxpayer, and I paid two cents of tax toward this program I think is illegal.â Youâd have hundreds of millions of people who could challenge any line item in the federal budget, and if you multiply that by how many line items someone might object to . . . Even though there is de minimis financial harm, and thatâs usually enough, weâve decided that isnât enough where the taxpayerâs interest is basically of the same nature as every other taxpayerâs interest in the matter. The analogy here is that we donât want 100 different people being able to unilaterally decide that A & C should be named. If 99 of them decide that maintaining anonymity is appropriate, and 1 disagrees, the odds of the 1 being correct are pretty low.]
If yes, it is hard to come up with a principled reason why Alice and Chloe canât be named now.
I expect no one was interested in writing something about Alice and/âor Chloe (A/âC), by name or otherwise, before Benâs post, and people only want to name them now because they think A/âC should face consequences for falsely (they believe) alleging abuse. Which is very close to retaliating against whistleblowers, and we should be very careful, which includes maybe accepting a rule that will have some false positives.
To take a different example, my non-professional understanding is it would normally be legal for an MA employer to report their employee to immigration authorities, but if the employer did this right after the employee had filed a complaint with the attorney generalâs office, even a false one, this is probably actually illegal retaliation. This will have the occasional false positive, where the employer really was going to report the employee anyway but canât prove it, but we accept that because avoiding the harms of retaliation is more important.
Jeffâactual âwhistleblowersâ make true and important allegations that withstand scrutiny and fact-checking. I agree that legit whistleblowers need the protection of anonymity.
But not all disgruntled ex-employees with a beef against their former bosses are whistleblowers in this sense. Many are pursuing their own retaliation strategies, often turning trivial or imagined slights into huge subjective moral outragesâand often getting credulous friends, family, journalists, or activists to support their cause and amplify their narrative.
Itâs true that most EAs had never heard of âAliceâ or âChloeâ, and didnât care about them, until they made public allegations against Nonlinear via Ben Paceâs post. And then, months later, many of us were dismayed and angry that many of their allegations turned out to be fabricated or exaggeratedâharming Nonlinear, wasting thousands of hours of our time, and creating schisms within out community.
So, arguably, we have a case here of two disgruntled ex-employees retaliating against a former employer. Why should their retaliation be protected by anonymity?
Conversely, when Kat Woods debunked many of the claims of Ben Pace (someone with much more power and influence in the EA/âRationalist community), why was she not considered a âwhistleblowerâ calling out his bullying and slander?
Yet again, the gender bias in âmoral typecastingâ becomes important, as I mentioned in a previous comment here.
So, arguably, we have a case here of two disgruntled ex-employees retaliating against a former employer. Why should their retaliation be protected by anonymity?
Highlighting that is an important crux (and one on which I have mixed feelings). Not all allegations of incorrect conduct rise to the level of âwhistleblowing.â A whistleblower brings alleged misconduct on a matter of public importance to light. We grant lots of protections in furtherance of that public interest, not out of regard for the whistleblowerâs private interests.
Is this a garden-variety dispute between an employer and two employees about terms of employment? Or is this a story about influential people allegedly using their power to mistreat two people who were in a vulnerable position which is of public import because it should update us on how much influence to allow those people?
In Australia people can be, and have been, prosecuted when they whistleblow on something commercially or otherwise sensitive (that was of major public importance!) by disclosing it publicly without completely exhausting internal whistleblowing processes. So even in cases of proper whistleblowing, countervailing factors can dominate in what the consequences are for whistleblower.
actual âwhistleblowersâ make true and important allegations that withstand scrutiny and fact-checking. I agree that legit whistleblowers need the protection of anonymity.
I think thatâs too strong? For example, under my amateur understanding of MAlaw I donât see anything about the anti-retailation provisions being conditional requiring a complaint to withstand scrutiny and fact-checking. And if this were changed to allow employers to retaliate in cases where employees claims were not sustained then I think weâd see, as a chilling effect, a decrease in employees raising true claims.
I agree that requiring that the claims be sustained would have a chilling effect. However, in many contexts, we donât extend protections to claims submitted in bad faith. For instance, we grant immunity from legal retaliation against people who file reports of child abuse . . . but that is usually conditioned on the allegations were made in good faith. If a reported individual can prove that the report was fabricated out of whole cloth, we donât shield the reporter from a defamation suit or other legal consequences.
Note that this is generally a subjective standardâif the reporter honestly believed the report was appropriate, we shield the reporter from liability. This doubtless allows some bad actors to slip through with immunity. However, we believe that is necessary to avoid reporters deciding not to report out of fear that someone will Monday-morning quarterback then and decide that reporting was objectively unreasonable.
In your example, I suspect that knowingly filing a false report with a state agency is a crime in MA (as it is with a federal agency at the federal level), so there is at least some potential enforcement mechanism for dealing with malicious lies.
I expect no one was interested in writing something about Alice and/âor Chloe (A/âC), by name or otherwise, before Benâs post [ . . . .]
(Correctly) surmising a lack of interest in writing a hypothetical expose about A/âC isnât quite the same thing as reaching a conclusion that the post shouldnât have been allowed to remain. However, I think there is a lot of overlap between the two; the reasons for lack of interest seem similar to the arguments for why the post shouldnât be allowed. So I think we are both somewhere vaguely near âthere would be no legitimate/âplausible reason for someone to write an expose about A/âC, unless one accepted that their whistleblowing activity made it legitimate.â
One interesting thing about this framing is that it raises the possibility that the whistleblowersâ identities are relevant to the decision. If a major figure in EA were going around telling malicious lies about other EAs, that would be (the subject of an appropriate post /â something people would be interested in writing about) independently of any specifically whistleblowing-retaliation angle.
One could stake out an anti-standing argument in which Nonlinear et al. would not be able to identify A/âC because we would be worried that vindictive or retaliatory desires were affecting their judgment, but an truly disinterested person (e.g., not friends of Nonlinear employees) could identify themâbecause they are likely to be acting from purer, less emotionally-invested motives (e.g., protection of the community from those they perceive as brazen liars). Iâm not endorsing that view, but it is interesting to ponder.
Continuing with your analogy, if you were a random person who found out about the employee filing with the AG independently of the employer, and you somehow were able to determine that the employee knowingly filed a false report out of (e.g.) racial animus, would it be OK for you to report the employee to the immigration authorities? [Asking ethically, not under MA law.]
Relatedly, your example relates to a situation with a significant power imbalance (employer/âemployee). It probably isnât illegal for me to report someone to the immigration authorities where my real motive is that they cheated on me, they cheated on my friend, etc. [I didnât specifically check that example.] So it seems that we often protect individuals exercising socially-important functions like whistleblowing from retaliation by some actors but not others.
My use of core expressive speech was inspired by âcore political speechâ in U.S. First Amendment doctrine. E.g., this article describing a hierarchy of protected speech. I meant that Forum speech may be more likely to be speech about the stuff that matters, discouragement of which (including by denying psuedonymity) poses particularly great harms. Probably âhigh-value speechâ would have been clearer here.
Solicitude is care or concern, so here I meant that we might particularly care about protecting Forum speech as opposed to other kinds of speech for some reason.
Hi Geoffrey, I think you raise a very reasonable point.
Thereâs some unfortunate timing at play here: 3â7 of the active mod teamâLizka, Toby, and JPâhave been away at a CEA retreat for the past ~week, and have thus mostly been offline. In my view, we would have ideally issued a proper update by now on the earlier notice: âFor the time being, please do not post personal information that would deanonymize Alice or Chloe.â
In lieu of that, Iâll instead publish one of my comments from the moderatorsâ Slack thread, along with some commentary. Iâm hoping that this shows some of our thinking and adds to the ongoing discussion here.[1] Iâm grateful to yourself, @Ivy Mazzola, @Jason, @Jeff Kaufman and others for helping push this conversation forward.
Starting context: Majority of moderators in agreement that our current policy on doxing, âRevealing someoneâs real name if they are anonymous on the Forum or elsewhere on the internet is prohibitedâ (link), should apply in the Alice+Chloe case. (And that it should apply whether or not Alice and/âor Chloe have exaggerated their allegations.)
Will (5 days ago)
I think what this comes down to for me is: If Kat Woodsâ Forum username was pseudonymous, would we have taken down Benâs post? (Or otherwise removed all references to Kat by her real name?)
If the answer to this is âyes,â then I donât think Alice+Chloe should be deanonymized.
If the answer to this is âno,â then I think Alice+Chloe should be deanonymized.[2] (Because if we go with âno,â then this would mean that itâs fair game for Kat to go write a post now that shares information on two past employees whom she believes have spread falsehoods about Nonlinear, using these employeesâ real names. Which is equivalent to deanonymizing Alice+Chloe.)
------
Iâm concerned about setting a precedent of first-mover advantage. Like, I imagine there are EAs with mutual grievances out there, and if we set a precedent of, âIf you strike first, anonymously, then you can name the person you bear a grievance against whilst granting yourself anonymity immunity,â then I think weâre laying the foundations for a pretty awful dynamic.
There was then commentary from a couple of the other moderators acknowledging that:
Under our current policies, it is indeed fair game for Kat to write a post sharing information on those former employees using their real names, as long as that post doesnât refer directly to Benâs one. Which does seem like a weird technicality.
The three broad paths forward seem to be: 1) âgive the first accuser the right to remain anonymous;â 2) âban anonymous allegations altogether;â 3) âstick with our current policies, notwithstanding that there is that technicality.[3]â
Thereâs since been the beginnings of a discussion on whether certain categories of anonymous allegationsânotably, employees sharing information on employersâcould be exceptions to the rule if the second path is taken.[4]
Closing context: At the time of me posting this comment, inter-moderator discussion on the issue is inconclusive.
I think it probably makes most sense to decide this at the policy level first, and then circle back to how to handle the Nonlinear case. Specifically, if we go with giving the first accuser anonymity rights at the policy level, then it follows that Alice and Chloe should remain anonymous. But if we decide that allegations should not be made anonymously, then weâll have to think of what to do with the Alice+Chloe allegations already out there.
[ETA: Like, since Alice and Chloe made their allegations expecting anonymity, I believe we should give their cases special treatment (if we go the route of not allowing anonymous allegations). This could just mean fully respecting their expectation. This could mean making difficult judgement calls in weighing up, for example, Aliceâs expectation for anonymity versus the magnitude of her exaggerations.]
And notwithstanding that this technicality may well put an accused party in an uncomfortable position, where they believe they have good reason for writing a response âsharing informationâ post, but they also know that doing so likely means receiving some backlash (for implicitly deanonymizing their accuser).
My personal view is that this rules out the third path: I donât think a fair set of policies would force someone into this kind of a lose-lose position.
In other words, one consideration is whether the benefit of allowing anonymous whistleblowing on the Forumâwhere the purpose of whistleblowing is in part to compensate for employer-employee asymmetryâoutweighs the cost of not having first-accuser-second-accuser symmetry.
I think what this comes down to for me is: If Kat Woodsâ Forum username was pseudonymous, would we have taken down Benâs post? (Or otherwise removed all references to Kat by her real name?)
If the answer to this is âyes,â then I donât think Alice+Chloe should be deanonymized.
I do not like the incentive structure that this would create if adopted. Kat did not get to look at this particular drama and decide whether she wanted it discussed under a real or pseudonymous username. Her decision point was when she created her forum account however many years ago, at a time when she had no idea that this kind of drama would erupt. If this position becomes policy, then it incentivizes every person, at the time that they create a forum account, to choose a pseudonym rather than use their real name, to avoid having any unforeseeable future drama publicly associated with their real name. I think this would be bad. People in a community canât build trust if they donât know the identities of the people they are building trust with.
A rule that you couldnât directly name people of moderate or greater prominence wouldnât work well anyway. People here are awfully clever, and Iâm sure one could easily write a whistleblowing piece on such a person that left very little doubt about their identity without actually saying their name or other unique identifiers. In fact, Iâm not sure if Benâs piece could have been effectively written without most of the Forum readership knowing who Alice and Chloe had worked for.
Willâthanks very much for sharing your views, and some of the discussion amongst the EA Forum moderators.
These are tricky issues, and Iâm glad to see that theyâre getting some serious attention, in terms of the relative costs, benefits, and risks of different possible politicies.
Iâm also concerned about âsetting a precedent of first-mover advantageâ. A blanket policy of first-mover (or first-accuser) anonymity would incentivize EAs to make lots of allegations before the people theyâre accusing could make counter-allegations. That seems likely to create massive problems, conflicts, toxicity, and schisms within EA.
I had a bunch of thoughts on this situation, enough that I wrote them up as a post. Unfortunately your response came out while I was writing and I didnât see it, but I think doesnât change much?
In addition to your three paths forward, I see a fourth one: you extend the policy to have the moderators (or another widely-trusted entity) make decisions on when there should be exceptions in cases like this, and write a bit about how youâll make those decisions.
There may be a fifth, which could be seen as a bit of a cop-out.
Itâs not clear to me whether the mods claim jurisdiction over deanonymizing conduct that doesnât happen on the Forum. I think the answer here is that claiming such jurisdiction here would be inappropriate.
As far as I know, it wouldnât violate the rules of X, Facebook, or most other sites to post that â[Real Names] have been spreading malicious lies about things that happened when they were Nonlinear employees.â It certainly would not violate the rules of a state or federal court to do that in a court complaint. The alleged harm of Alice and Chloe spreading malicious lies about Kat, Emerson, and Nonlinear existed off-the-Forum prior to anything being published on the Forum. I donât see why Benâs act of including those allegations in a Forum post creates off-Forum obligations for Nonlinear et al. (or anyone else) that did not exist prior to Benâs post. Alice and Chloe, and people in similar situations, have to accept that many fora exist that do not have norms against this kind of conduct.[1]
If there is no jurisdiction over off-Forum naming here, it seems that the people who want Alice and Chloe named can do so in other places, and everyone who wants to know will know soon enough. If thatâs the case, Iâm not sure whetherâat least in these circumstances [2] -- Jeffâs suggestion offers enough added value to justify the rather significant costs of the mods adjudicating this particular matter in a reasonably thorough manner. If the names can be plastered all over X and Facebook, and ~everyone who cares to know will find out that way, does it make a huge amount of difference whether or not the names are also on the Forum? Under those circumstances, declining to adjudicate because the question would be of limited practical importance would be justifiable.
I express no opinion as to whether the mods could legitimately exercise broader jurisdiction over attempting to disclose the identity of a Forum poster, where the only relevant conduct was on-Forum.
This isnât about the Forum mods as as representatives of the Forum, but instead as the most obvious trusted community members (possibly in consultation with CH) to make a decision.
What centralized adjudication avoids is each person having to make their own judgment about whether deanonymization is appropriate in a given circumstance. Letâs say NL starts posting the real names on Twitter: should I think poorly of them for breaking an important norm or is this an exception? Is that an unreasonable unilateral escalation of this dispute? Should I pressure them not to do this?
That approach certainly does offer some significant advantages, but I think itâs a lot harder to pull off. Willâs three options, the narrower version of mod discretion (which is limited to whether A/âC can be named on the Forum), and my fifth option (declining to allow in this case because if people decide to name, everyone will find out whether itâs on the Forum or not) are all open to the mods because they are mods.
The possibility of a centralized adjudication that is recognized as binding in all places requires outside buy-in. I think it needs either (1) the consent of every party directly in interest or (2) the consent of Nonlinear, broad community support, and the centralized adjudicatorâs willingness to either release the names themselves or allow widespread burner accounts naming them.
Option (1) is basically arbitration on the consent of the parties; they would be free to choose the mods, Qualy, a poll, or a coin flip. Alice and Chloe would consent to being named if the arbitrators chose, and Nonlinear would agree to not name if the arbitrators ruled against them.[1] If the arbitrators rule for naming, no one should judge Nonlinear because it would have named Alice and Chloe with their consent. If they rule against naming and Nonlinear did it anyway, everyone should judge them for breaking their agreement. And thereâs a strong argument to me that we bystanders should honor the decision of those directly involved on a resolution.
But reaching an agreement to arbitrate may be challenging. A rational party would not consent to arbitrate unless it concluded its interests were expected to be better off under arbitration than the counterfactual. Settlements can be mutually beneficial, but I am not yet convinced arbitration would be in Alice and Chloeâs interests. So long as a substantial fraction of the community would judge Nonlinear for naming, it probably will not do so. So the status quo for Alice and Chloe would be a win vs. an uncertain future in arbitration.
The other, less certain option is that Nonlinear and the significant majority of the community consented to abide by the arbitration result. Even here, there is a risk that the arbitration process may become a no-win scenario for Nonlinear. If enough community members reserve their right to independently and adversely judge Nonlinear for naming, then it is in a pickle even if it âwinsâ the arbitration.
A possible workaround might be either that the arbitration panel itself will release the names if it rules against Alice and Chloe, or that it will allow untraceable anonymous posters to flood the Forum with their names. In other words, if Alice and Chloe do not consent, and there is a contingent of anti-namers in the community, then any blowback for releasing needs to fall on someone other than Nonlinear.
The possibility of a centralized adjudication that is recognized as binding in all places requires outside buy-in.
I think you might be thinking too formally? We sometimes have things that work because we decide to respect an authority that doesnât have any formal power. If you make a film you donât have to submit it to the MPAA to get a rating, and if you run a theater you donât have to follow MPAA ratings in deciding whether someone is mature enough to be let into an R-rated movie, but everyone just goes along with the system.
Iâm imagining that the Forum mods would make a decision for the Forum, and then weâd just go along with it voluntarily even off the Forum, as long as they kept making reasonable decisions.
Iâm not seeing any real consensus on what standard to apply for deanonymizing someone. I think a voluntary deference model is much easier when such a consensus exists. If youâre on board with the basic decision standard, itâs easier to defer even when you disagree with the application in a specific case. In sports, the referees usually get the call right, and errors are evenly distributed between your team and your opponents. But if you fundamentally disagree with the decision standard, the calls will go systematically against your viewpoint. Thatâs much harder to defer to, and people obviously have very strong feelings on either side.
I donât think the MPAA is a great analog here. Iâd submit that the MPAA has designed its system carefully in light of the wholly advisory nature of its rulings. Placing things on a five-point continuum helps. I think only a small fraction of users would disagree more than one rating up/âdown from where the MPAA lands. So rarely would an end user completely disagree with the MPAA outcome. Where an end user knows that the MPAA grades more harshly/âleniently than they do, the user can mentally adjust accordingly (as they might when they learn so many Harvard College students get 4.0s?)
And itâs easy for end users to practically opt out of the MPAA system without any real social sanction; if a theater owner decides to admit ten-year olds to R-rated movies with a signed parental consent, that is really none of my business. If a parent decides to take a seven-year old to one, that is also none of my business as long as the child is non-disruptive. The MPAA system is resilient to 20-30% of the population opting out, if they so chose.[1]
So I donât think the features that make the MPAA system workable as a voluntary-deference system are likely to transfer over well to this context.
Itâs harder for filmmakers to opt outâbut they also got a lot out of the system, too. Mild self-regulation is preferable to government regulation, especially back in days when the First Amendment was not enforced with the same rigor it is today.
Jeffâthanks very much for sharing the link to that post. I encourage others to read itâitâs fairly short. It nicely sets out some of the difficulties around anonymity, doxxing, accusations, counter-accusations, etc.
I canât offer any brilliant solutions to these issues, but I am glad to see that the risks of false or exaggerated allegations are getting some serious attention.
I wouldnât classify Benâs post as containing fully anonymous allegations. There was a named community member who implicitly vouched for the allegations having enough substance to lay before the Forum community. That means there was someone in a position to accept social and legal fallout if the decision to post those allegations is proven to have been foolhardy. That seems to be a substantial safeguard against the posting of spurious nonsense.
Maybe having such a person identified didnât work out here, but I think itâs worth distinguishing between this case and a truly anonymous situation (e.g., burner account registered with throwaway account doing business via Tor, with low likelihood that even the legal system could identify the actual poster for imposition of consequences).
And notwithstanding that this technicality might put the accused party in an uncomfortable position, where they believe they have good reason for writing a response âsharing informationâ post, but they also know that doing so will likely make them the target of some severe backlash for implicitly deanonymizing their accuser.
That could be a feature rather than a bug for reasons similar to those described above. Deanonymizing someone who claims to be a whistleblower is a big dealâand arguably we should require an identified poster to accept the potential social and legal fallout if that decision wasnât warranted, as a way of discouraging inappropriate deanonymization.
A quick reminder that moderators have asked, at least for the time being, to please not post personal information that would deanonymize Alice or Chloe.
Lorenzoâyes, Iâm complying with that request.
Iâm just puzzled about the apparent double standard where the first people to make allegations enjoy privacy & anonymity (even if their allegations seem to be largely false or exaggerated), but the people theyâre accusing donât enjoy the same privilege.
I agree that the Forumâs rules and norms on privacy protection are confused. A few observations:
(1) Suppose a universe in which the first post on this topic had been from from Nonlinear, and had accused Alice and Chloe (by their real names) of a pattern of mendaciously spreading lies about Nonlinear. Would that post have been allowed to stay up? If yes, it is hard to come up with a principled reason why Alice and Chloe canât be named now.
If no, we would need to think about why this hypothetical post would have been disallowed. The best argument I came with would be that Alice and Chloe are, as far as I know, people with no real prominence/âinfluence/âpower (âPIPâ) within or without EA. Under this argument, there is a greater public interest in the actions of those with PIP, and accepting a role of PIP necessarily means sacrificing some of the privacy rights that non-PIPs get.
(2) Another possibility involves the idea of standing. Under this theory, Alice and Chloe had standing to name people at Nonlinear because they were the ones who allegedly experienced harm. Ben had derivative standing because Alice and Chloe had given him permission to share their stories on the Forum. Under this theory, Nonlinear (and individuals named in Benâs post) would have standing to name Alice and Chloe as the parties allegedly harmed by their conduct. The rest of us would not have standing. [Edit to add footnote here.[1]]
Maybe the aggrieved party isnât the best imaginable arbiter of whether someone should be unmasked, but itâs probably better than concluding that each and every one of us get to make that decision. That could more easily lead to a situation in which 99% of us agree that the names shouldnât be disclosed, and the 1% gets to decide. Of course, that could happen under a standing theory as well, but the 1% minority opinion has to coincide with the list of parties with standing. That at least narrows the problem.
(3) This situation involves a mix of Forum and non-Forum conduct. I sense that the norm for not doxxing people is stronger for Forum conduct (or at least for conduct involving similar sorts of speech) than from non-Forum conduct. This may reflect a difference in the norms of Internet communities vs. physically-grounded communities, or a sense that Forum speech is often core expressive speech or otherwise entitled to extra solicitude.
In any event, based on the fact that a number of people seem to know their identities, it sounds like the tale of Alice and Chloe was not a closely-kept secret before Benâs post went up. Presumably this information came, directly or indirectly, from Alice and Chloe. In other words, there has been potentially relevant non-Forum conduct that did not magically morph into Forum conduct merely because Ben decided to write a post sharing the same information.
That raises the following hypothetical. Letâs say that Dan allegedly made some homophobic, racist, sexist, and/âor otherwise offensive comments.[2] Does the location in which the comments were made change whether he can be identified. For example, is it OK to write a Forum post identifying Dan by name if those comments were posted on the Forum under a pseudonym? What if they were said at an EAG afterparty? In a non-EA space on Reddit? With a mask on in public (by someone who recognized Danâs voice)? If the answers are not the same, why is that the case? I donât claim to know the answers, by the way; this isnât a trick question.
(4) I do not at present have a clear opinion on whether naming Alice, Chloe, or both should be allowed on the Forum. I do have an opinion that there needs to be a clear, logical set of rules or at least principlesâideally, laid out in advance as far as that is practical. I donât like the idea of apparently ad hoc decisions about who gets privacy protections and who does not.
[Added footnote: I should note that standing, which is used as a legal metaphor here, does not mean âis none of your business.â We often impose standing requirements to make sure the person making the decision has appropriate incentives, context, etc.
Here, the various people who A & C have allegedly spread lies about are in a much better position than I am to know the relevant facts. If they have not concluded that disclosure of A & Câs names is warranted, itâs not clear why Iâor any other reader of the relevant postsâwould do a better job as decisionmaker.
Another reason for limited standing is practicality. For instance, in US law, we donât usually allow suits where the alleged harm is âIâm a taxpayer, and I paid two cents of tax toward this program I think is illegal.â Youâd have hundreds of millions of people who could challenge any line item in the federal budget, and if you multiply that by how many line items someone might object to . . . Even though there is de minimis financial harm, and thatâs usually enough, weâve decided that isnât enough where the taxpayerâs interest is basically of the same nature as every other taxpayerâs interest in the matter. The analogy here is that we donât want 100 different people being able to unilaterally decide that A & C should be named. If 99 of them decide that maintaining anonymity is appropriate, and 1 disagrees, the odds of the 1 being correct are pretty low.]
This is not to equate Danâs hypothetical conduct to what Alice and Chloe are alleged to have done.
I expect no one was interested in writing something about Alice and/âor Chloe (A/âC), by name or otherwise, before Benâs post, and people only want to name them now because they think A/âC should face consequences for falsely (they believe) alleging abuse. Which is very close to retaliating against whistleblowers, and we should be very careful, which includes maybe accepting a rule that will have some false positives.
To take a different example, my non-professional understanding is it would normally be legal for an MA employer to report their employee to immigration authorities, but if the employer did this right after the employee had filed a complaint with the attorney generalâs office, even a false one, this is probably actually illegal retaliation. This will have the occasional false positive, where the employer really was going to report the employee anyway but canât prove it, but we accept that because avoiding the harms of retaliation is more important.
Jeffâactual âwhistleblowersâ make true and important allegations that withstand scrutiny and fact-checking. I agree that legit whistleblowers need the protection of anonymity.
But not all disgruntled ex-employees with a beef against their former bosses are whistleblowers in this sense. Many are pursuing their own retaliation strategies, often turning trivial or imagined slights into huge subjective moral outragesâand often getting credulous friends, family, journalists, or activists to support their cause and amplify their narrative.
Itâs true that most EAs had never heard of âAliceâ or âChloeâ, and didnât care about them, until they made public allegations against Nonlinear via Ben Paceâs post. And then, months later, many of us were dismayed and angry that many of their allegations turned out to be fabricated or exaggeratedâharming Nonlinear, wasting thousands of hours of our time, and creating schisms within out community.
So, arguably, we have a case here of two disgruntled ex-employees retaliating against a former employer. Why should their retaliation be protected by anonymity?
Conversely, when Kat Woods debunked many of the claims of Ben Pace (someone with much more power and influence in the EA/âRationalist community), why was she not considered a âwhistleblowerâ calling out his bullying and slander?
Yet again, the gender bias in âmoral typecastingâ becomes important, as I mentioned in a previous comment here.
Highlighting that is an important crux (and one on which I have mixed feelings). Not all allegations of incorrect conduct rise to the level of âwhistleblowing.â A whistleblower brings alleged misconduct on a matter of public importance to light. We grant lots of protections in furtherance of that public interest, not out of regard for the whistleblowerâs private interests.
Is this a garden-variety dispute between an employer and two employees about terms of employment? Or is this a story about influential people allegedly using their power to mistreat two people who were in a vulnerable position which is of public import because it should update us on how much influence to allow those people?
In Australia people can be, and have been, prosecuted when they whistleblow on something commercially or otherwise sensitive (that was of major public importance!) by disclosing it publicly without completely exhausting internal whistleblowing processes. So even in cases of proper whistleblowing, countervailing factors can dominate in what the consequences are for whistleblower.
I think thatâs too strong? For example, under my amateur understanding of MA law I donât see anything about the anti-retailation provisions being conditional requiring a complaint to withstand scrutiny and fact-checking. And if this were changed to allow employers to retaliate in cases where employees claims were not sustained then I think weâd see, as a chilling effect, a decrease in employees raising true claims.
I agree that requiring that the claims be sustained would have a chilling effect. However, in many contexts, we donât extend protections to claims submitted in bad faith. For instance, we grant immunity from legal retaliation against people who file reports of child abuse . . . but that is usually conditioned on the allegations were made in good faith. If a reported individual can prove that the report was fabricated out of whole cloth, we donât shield the reporter from a defamation suit or other legal consequences.
Note that this is generally a subjective standardâif the reporter honestly believed the report was appropriate, we shield the reporter from liability. This doubtless allows some bad actors to slip through with immunity. However, we believe that is necessary to avoid reporters deciding not to report out of fear that someone will Monday-morning quarterback then and decide that reporting was objectively unreasonable.
In your example, I suspect that knowingly filing a false report with a state agency is a crime in MA (as it is with a federal agency at the federal level), so there is at least some potential enforcement mechanism for dealing with malicious lies.
(Correctly) surmising a lack of interest in writing a hypothetical expose about A/âC isnât quite the same thing as reaching a conclusion that the post shouldnât have been allowed to remain. However, I think there is a lot of overlap between the two; the reasons for lack of interest seem similar to the arguments for why the post shouldnât be allowed. So I think we are both somewhere vaguely near âthere would be no legitimate/âplausible reason for someone to write an expose about A/âC, unless one accepted that their whistleblowing activity made it legitimate.â
One interesting thing about this framing is that it raises the possibility that the whistleblowersâ identities are relevant to the decision. If a major figure in EA were going around telling malicious lies about other EAs, that would be (the subject of an appropriate post /â something people would be interested in writing about) independently of any specifically whistleblowing-retaliation angle.
One could stake out an anti-standing argument in which Nonlinear et al. would not be able to identify A/âC because we would be worried that vindictive or retaliatory desires were affecting their judgment, but an truly disinterested person (e.g., not friends of Nonlinear employees) could identify themâbecause they are likely to be acting from purer, less emotionally-invested motives (e.g., protection of the community from those they perceive as brazen liars). Iâm not endorsing that view, but it is interesting to ponder.
Continuing with your analogy, if you were a random person who found out about the employee filing with the AG independently of the employer, and you somehow were able to determine that the employee knowingly filed a false report out of (e.g.) racial animus, would it be OK for you to report the employee to the immigration authorities? [Asking ethically, not under MA law.]
Relatedly, your example relates to a situation with a significant power imbalance (employer/âemployee). It probably isnât illegal for me to report someone to the immigration authorities where my real motive is that they cheated on me, they cheated on my friend, etc. [I didnât specifically check that example.] So it seems that we often protect individuals exercising socially-important functions like whistleblowing from retaliation by some actors but not others.
All good points! Iâm quite conflicted here.
Could you explain what â core expressive speechâ and âextra solicitudeâ are?
My use of core expressive speech was inspired by âcore political speechâ in U.S. First Amendment doctrine. E.g., this article describing a hierarchy of protected speech. I meant that Forum speech may be more likely to be speech about the stuff that matters, discouragement of which (including by denying psuedonymity) poses particularly great harms. Probably âhigh-value speechâ would have been clearer here.
Solicitude is care or concern, so here I meant that we might particularly care about protecting Forum speech as opposed to other kinds of speech for some reason.
Writing in a personal capacity.
Hi Geoffrey, I think you raise a very reasonable point.
Thereâs some unfortunate timing at play here: 3â7 of the active mod teamâLizka, Toby, and JPâhave been away at a CEA retreat for the past ~week, and have thus mostly been offline. In my view, we would have ideally issued a proper update by now on the earlier notice: âFor the time being, please do not post personal information that would deanonymize Alice or Chloe.â
In lieu of that, Iâll instead publish one of my comments from the moderatorsâ Slack thread, along with some commentary. Iâm hoping that this shows some of our thinking and adds to the ongoing discussion here.[1] Iâm grateful to yourself, @Ivy Mazzola, @Jason, @Jeff Kaufman and others for helping push this conversation forward.
Starting context: Majority of moderators in agreement that our current policy on doxing, âRevealing someoneâs real name if they are anonymous on the Forum or elsewhere on the internet is prohibitedâ (link), should apply in the Alice+Chloe case. (And that it should apply whether or not Alice and/âor Chloe have exaggerated their allegations.)
There was then commentary from a couple of the other moderators acknowledging that:
Under our current policies, it is indeed fair game for Kat to write a post sharing information on those former employees using their real names, as long as that post doesnât refer directly to Benâs one. Which does seem like a weird technicality.
The three broad paths forward seem to be: 1) âgive the first accuser the right to remain anonymous;â 2) âban anonymous allegations altogether;â 3) âstick with our current policies, notwithstanding that there is that technicality.[3]â
Thereâs since been the beginnings of a discussion on whether certain categories of anonymous allegationsânotably, employees sharing information on employersâcould be exceptions to the rule if the second path is taken.[4]
Closing context: At the time of me posting this comment, inter-moderator discussion on the issue is inconclusive.
The thinking shown will of course be skewed towards my personal view, over the views of other moderators.
A later comment of mine, which ties in here:
And notwithstanding that this technicality may well put an accused party in an uncomfortable position, where they believe they have good reason for writing a response âsharing informationâ post, but they also know that doing so likely means receiving some backlash (for implicitly deanonymizing their accuser).
My personal view is that this rules out the third path: I donât think a fair set of policies would force someone into this kind of a lose-lose position.
In other words, one consideration is whether the benefit of allowing anonymous whistleblowing on the Forumâwhere the purpose of whistleblowing is in part to compensate for employer-employee asymmetryâoutweighs the cost of not having first-accuser-second-accuser symmetry.
I do not like the incentive structure that this would create if adopted. Kat did not get to look at this particular drama and decide whether she wanted it discussed under a real or pseudonymous username. Her decision point was when she created her forum account however many years ago, at a time when she had no idea that this kind of drama would erupt. If this position becomes policy, then it incentivizes every person, at the time that they create a forum account, to choose a pseudonym rather than use their real name, to avoid having any unforeseeable future drama publicly associated with their real name. I think this would be bad. People in a community canât build trust if they donât know the identities of the people they are building trust with.
A rule that you couldnât directly name people of moderate or greater prominence wouldnât work well anyway. People here are awfully clever, and Iâm sure one could easily write a whistleblowing piece on such a person that left very little doubt about their identity without actually saying their name or other unique identifiers. In fact, Iâm not sure if Benâs piece could have been effectively written without most of the Forum readership knowing who Alice and Chloe had worked for.
Willâthanks very much for sharing your views, and some of the discussion amongst the EA Forum moderators.
These are tricky issues, and Iâm glad to see that theyâre getting some serious attention, in terms of the relative costs, benefits, and risks of different possible politicies.
Iâm also concerned about âsetting a precedent of first-mover advantageâ. A blanket policy of first-mover (or first-accuser) anonymity would incentivize EAs to make lots of allegations before the people theyâre accusing could make counter-allegations. That seems likely to create massive problems, conflicts, toxicity, and schisms within EA.
Thanks for sharing this!
I had a bunch of thoughts on this situation, enough that I wrote them up as a post. Unfortunately your response came out while I was writing and I didnât see it, but I think doesnât change much?
In addition to your three paths forward, I see a fourth one: you extend the policy to have the moderators (or another widely-trusted entity) make decisions on when there should be exceptions in cases like this, and write a bit about how youâll make those decisions.
There may be a fifth, which could be seen as a bit of a cop-out.
Itâs not clear to me whether the mods claim jurisdiction over deanonymizing conduct that doesnât happen on the Forum. I think the answer here is that claiming such jurisdiction here would be inappropriate.
As far as I know, it wouldnât violate the rules of X, Facebook, or most other sites to post that â[Real Names] have been spreading malicious lies about things that happened when they were Nonlinear employees.â It certainly would not violate the rules of a state or federal court to do that in a court complaint. The alleged harm of Alice and Chloe spreading malicious lies about Kat, Emerson, and Nonlinear existed off-the-Forum prior to anything being published on the Forum. I donât see why Benâs act of including those allegations in a Forum post creates off-Forum obligations for Nonlinear et al. (or anyone else) that did not exist prior to Benâs post. Alice and Chloe, and people in similar situations, have to accept that many fora exist that do not have norms against this kind of conduct.[1]
If there is no jurisdiction over off-Forum naming here, it seems that the people who want Alice and Chloe named can do so in other places, and everyone who wants to know will know soon enough. If thatâs the case, Iâm not sure whetherâat least in these circumstances [2] -- Jeffâs suggestion offers enough added value to justify the rather significant costs of the mods adjudicating this particular matter in a reasonably thorough manner. If the names can be plastered all over X and Facebook, and ~everyone who cares to know will find out that way, does it make a huge amount of difference whether or not the names are also on the Forum? Under those circumstances, declining to adjudicate because the question would be of limited practical importance would be justifiable.
I express no opinion as to whether the mods could legitimately exercise broader jurisdiction over attempting to disclose the identity of a Forum poster, where the only relevant conduct was on-Forum.
That is, enough people who know of their identity appear to be motivated to share it.
This isnât about the Forum mods as as representatives of the Forum, but instead as the most obvious trusted community members (possibly in consultation with CH) to make a decision.
What centralized adjudication avoids is each person having to make their own judgment about whether deanonymization is appropriate in a given circumstance. Letâs say NL starts posting the real names on Twitter: should I think poorly of them for breaking an important norm or is this an exception? Is that an unreasonable unilateral escalation of this dispute? Should I pressure them not to do this?
That approach certainly does offer some significant advantages, but I think itâs a lot harder to pull off. Willâs three options, the narrower version of mod discretion (which is limited to whether A/âC can be named on the Forum), and my fifth option (declining to allow in this case because if people decide to name, everyone will find out whether itâs on the Forum or not) are all open to the mods because they are mods.
The possibility of a centralized adjudication that is recognized as binding in all places requires outside buy-in. I think it needs either (1) the consent of every party directly in interest or (2) the consent of Nonlinear, broad community support, and the centralized adjudicatorâs willingness to either release the names themselves or allow widespread burner accounts naming them.
Option (1) is basically arbitration on the consent of the parties; they would be free to choose the mods, Qualy, a poll, or a coin flip. Alice and Chloe would consent to being named if the arbitrators chose, and Nonlinear would agree to not name if the arbitrators ruled against them.[1] If the arbitrators rule for naming, no one should judge Nonlinear because it would have named Alice and Chloe with their consent. If they rule against naming and Nonlinear did it anyway, everyone should judge them for breaking their agreement. And thereâs a strong argument to me that we bystanders should honor the decision of those directly involved on a resolution.
But reaching an agreement to arbitrate may be challenging. A rational party would not consent to arbitrate unless it concluded its interests were expected to be better off under arbitration than the counterfactual. Settlements can be mutually beneficial, but I am not yet convinced arbitration would be in Alice and Chloeâs interests. So long as a substantial fraction of the community would judge Nonlinear for naming, it probably will not do so. So the status quo for Alice and Chloe would be a win vs. an uncertain future in arbitration.
The other, less certain option is that Nonlinear and the significant majority of the community consented to abide by the arbitration result. Even here, there is a risk that the arbitration process may become a no-win scenario for Nonlinear. If enough community members reserve their right to independently and adversely judge Nonlinear for naming, then it is in a pickle even if it âwinsâ the arbitration.
A possible workaround might be either that the arbitration panel itself will release the names if it rules against Alice and Chloe, or that it will allow untraceable anonymous posters to flood the Forum with their names. In other words, if Alice and Chloe do not consent, and there is a contingent of anti-namers in the community, then any blowback for releasing needs to fall on someone other than Nonlinear.
One complexity is that, to the extent that Alice and/âor Chloe allegedly slandered other people, there are other potential parties in interest.
I think you might be thinking too formally? We sometimes have things that work because we decide to respect an authority that doesnât have any formal power. If you make a film you donât have to submit it to the MPAA to get a rating, and if you run a theater you donât have to follow MPAA ratings in deciding whether someone is mature enough to be let into an R-rated movie, but everyone just goes along with the system.
Iâm imagining that the Forum mods would make a decision for the Forum, and then weâd just go along with it voluntarily even off the Forum, as long as they kept making reasonable decisions.
Iâm not seeing any real consensus on what standard to apply for deanonymizing someone. I think a voluntary deference model is much easier when such a consensus exists. If youâre on board with the basic decision standard, itâs easier to defer even when you disagree with the application in a specific case. In sports, the referees usually get the call right, and errors are evenly distributed between your team and your opponents. But if you fundamentally disagree with the decision standard, the calls will go systematically against your viewpoint. Thatâs much harder to defer to, and people obviously have very strong feelings on either side.
I donât think the MPAA is a great analog here. Iâd submit that the MPAA has designed its system carefully in light of the wholly advisory nature of its rulings. Placing things on a five-point continuum helps. I think only a small fraction of users would disagree more than one rating up/âdown from where the MPAA lands. So rarely would an end user completely disagree with the MPAA outcome. Where an end user knows that the MPAA grades more harshly/âleniently than they do, the user can mentally adjust accordingly (as they might when they learn so many Harvard College students get 4.0s?)
And itâs easy for end users to practically opt out of the MPAA system without any real social sanction; if a theater owner decides to admit ten-year olds to R-rated movies with a signed parental consent, that is really none of my business. If a parent decides to take a seven-year old to one, that is also none of my business as long as the child is non-disruptive. The MPAA system is resilient to 20-30% of the population opting out, if they so chose.[1]
So I donât think the features that make the MPAA system workable as a voluntary-deference system are likely to transfer over well to this context.
Itâs harder for filmmakers to opt outâbut they also got a lot out of the system, too. Mild self-regulation is preferable to government regulation, especially back in days when the First Amendment was not enforced with the same rigor it is today.
Jeffâthanks very much for sharing the link to that post. I encourage others to read itâitâs fairly short. It nicely sets out some of the difficulties around anonymity, doxxing, accusations, counter-accusations, etc.
I canât offer any brilliant solutions to these issues, but I am glad to see that the risks of false or exaggerated allegations are getting some serious attention.
I wouldnât classify Benâs post as containing fully anonymous allegations. There was a named community member who implicitly vouched for the allegations having enough substance to lay before the Forum community. That means there was someone in a position to accept social and legal fallout if the decision to post those allegations is proven to have been foolhardy. That seems to be a substantial safeguard against the posting of spurious nonsense.
Maybe having such a person identified didnât work out here, but I think itâs worth distinguishing between this case and a truly anonymous situation (e.g., burner account registered with throwaway account doing business via Tor, with low likelihood that even the legal system could identify the actual poster for imposition of consequences).
That could be a feature rather than a bug for reasons similar to those described above. Deanonymizing someone who claims to be a whistleblower is a big dealâand arguably we should require an identified poster to accept the potential social and legal fallout if that decision wasnât warranted, as a way of discouraging inappropriate deanonymization.