I myself doubt that any disclosure etc is fair. The ACM is not a trial court that protects individual rights - it should not be making findings about people as if it is (if the ACM wants to set up a Tribunal akin to the Medical Profession, it can do so, but that’s a world of difference to what it currently has). Such an exercise is likely to breach GDPR rights, because its not processing people’s data fairly - blacklists made by non-judicial (or even non-quazi-judicial) bodies of this nature are not exactly the type of thing that have anything but an unedifying history (e.g. with construction workers in the UK). Moreover, it might create a lot of oddities - different institutions will be required to treat that information differently. Its one thing for the ACM - purely through its administrative staff - to stop people registering for a conference. Its another to make it available to a fair number of academics, who as academics do, will gossip extensively.
As it stands, this policy does not speak to this situation. Thank you for bringing it to our attention.
On behalf of SIGPLAN CARES:
-
I think it would be helpful if the article more clearly specifies the differences between the current policy and the new proposed one. The authors are so familiar with it that they may not realise that random ACM members may have little or no idea of the current policy. A table with two columns (current policy, proposed new policy) and a row for each significant difference, would be helpful.
-
I think it would really help if the policy spelled out the principles that it seeks to embody, thus helping those who are making the case-by-case decisions that the body of the text says may happen. For example:
-
Disclosure is only ever made for a purpose, not simply to blame or shame. Purposes may include:
-
Policy in action. Giving ACM members concrete examples of its policies on harassment being enacted, in a timely and effective manner. (This may in part be done through aggregated statistics, such as those in the article, and an annual report.)
-
Encouraging victims. Thereby encouraging those who are the subject of harassment to come forward, by giving them evidence that their claims will be taken seriously and acted on.
-
Innocent until proven guilty. The policy recognises that the mere accusation of mis-conduct, let alone an upheld complaint, can and probably will have a career-changing impact on the respondent. They may be shunned, excluded, and despised. If the accusation proves false, these outcomes would be deeply unjust. Even if the accusation proves true, they make it harder for the respondent to improve their behaviour and rehabilitate.
-
Prevention of future harm. It is unacceptable for a known predator to be allowed to continue abuse unchecked within ACM. It follows that those with a need to know (program chairs, etc) should know of respondents where ACM has determined guilt.
-
-
Disclosure is only ever made for a purpose, not simply to blame or shame. Purposes may include:
(These purposes are implicit in the accompanying article, but I’d love to see them listed explicitly in the policy.)
-
Disclosure should be influenced by evidence of a pattern of behaviour. A single event may be a lapse; the danger of future harm is less, and rehabilitation seems more likely. On the other hand a pattern of repeated behaviour suggests something more serious: repeated harm suggests future harm; and rehabilitation seems less plausible.
-
Need to know. Draft 2 says who will be informed in a way that discloses the identity of the respondent. It neglects to say who will not be. Notably, as I understand it, there are no circumstances in which ACM would publicly report the name of the respondent or penalties – except if the matter had become public knowledge anyway. (The case of Tao Li seems like a counter-example?) I think it would be helpful to articulate the principle here.
Thank you to ACM for creating this disclosure policy proposal and to the community members who recognized the need for more disclosure and pushed for the creation of this proposal. I am in favor of the proposed policy. I, too, appreciate the distinction being made between minor and severe violations.
Thanks to the committee for their work and service. We have faced a number of challenges in SIGARCH regarding violations of professional conduct in our community in recent years. These challenges have led to a number of questions which we communicated publicly in a blog titled “Questions about policies in the wake of JIC”. We appreciate that ACM is moving forward towards addressing such questions.
I am strongly in favor of disclosure for serious violations to make sure that we protect the younger members of the community. I also would like to second the comment by @dchen that the benefits of pursuing credible allegations of violation and disclosure of serious violations for the community outweigh the risks of eventual legal threats to ACM.
Babak Falsafi
SIGARCH Chair
I appreciate the effort the ACM has put in to reconsider and update their policies around disclosure. These policies are tricky to get right and I suspect there will always be room for improvement. In that context, this draft and attempt represents a significant improvement over the status quo of no policy and zero disclosure. Recent experience has revealed many shortcomings in that approach, and I support this effort to remedy it.
Martha Kim
Columbia University
Member, SIGARCH EC
Thank you for putting together this thoughtful draft. It succeeds at balancing the need to disclose more broadly in cases where it may avert further harm while unnecessarily minimizing the airing of details otherwise.
As I understand it, the document draws a line between “minor” and “more serious than minor,” confines additional reporting considerations to the latter, and within that category suggests that a determination be made based primarily on potential harm if not disclosed. I find this general approach judicious.
However, the document does not currently describe what violations are “minor” or “more serious than minor.” To the extent that “more serious than minor” probably encompasses a broad spectrum of violations, it would be helpful to provide clarity. Specifically, there isn’t a precise mapping between these two categories and the levels described in the different policies for handling violations currently in place (all of which will presumably remain in effect). For example, the penalties for publication violations currently are categorized as Incidental, Low-level, Moderate, Significant, and Severe.
Furthermore, the document does not outline how ACM will build guidance on what specific violations warrant additional disclosures. I think that it is reasonable to assume that, at least initially, decisions on additional disclosure may be taken on a case-by-case basis. However, it would be important to outline a framework that provides uniformity and a sense of precedent. I suggest that a single committee, panel, or other similar (new or existing) body within ACM channel all requests for additional disclosures that fall under this category, be responsible for issuing a ruling, and develop guidance over time.
On the issue of minimal reporting, as I’ve written elsewhere in this forum, I disagree that witnesses should always be informed of the nature of the violation. Not all witnesses are victims, and not all of them may even be aware of the underlying complaint.
My context: I’m a professor at Cornell University and member of the the SIGARCH Board of Directors.
José Martínez
Current policy is to not disclose anything in the cases involving violations of the Code of Ethics or the Policy Against Harassment. Violations of Publications Policies can disclose more (e.g. a plagiarized paper is marked as such in the Digital Library and in certain cases violators superiors are notified).
Also, thank you for the additional remarks that are clarifying.
It starts to feel that the push for Draft 2 may be over-influenced by some advocators in the SIGARCH community. We need to keep in mind that any adopted policy would be applied to all ACM communities with 100k members across a hundred countries as well as non-members who also engage with ACM events. I am deeply concerned about the current proposal when viewed more broadly:
-
What if violators are students? In the university setting, even if a student commits a severe academic misconduct in a course that warrants suspension, we don’t disclose the misconduct to instructors in other courses that the student may enroll in, and we certainly don’t disclose to other universities (for legal and moral reasons). Why is it suddenly okay to disregard these principles and ruin students’ future?
-
If the violator is a professor whose identity is disclosed, would it be unfair to former and other current students of the professor when they seek faculty jobs? Those students did nothing wrong but may likely be disadvantaged, as people would naturally relate them to the professor.
-
Many people learn from mistakes, especially after a serious one with severe penalty. Draft 2 proposes to broaden disclosure regardless, which is simply a speculation. Disclosure should be made only if there is creditable evidence suggesting that similar misconducts will likely reoccur.
-
I feel that the moment we start to compare benefits vs risks, we may have been on the wrong path. Violators are still humans and have human rights. They are protected by constitution which is not up for discussion on whether sacrificing the rights a little can be justified by greater benefits. Taking defamation mentioned in the report for example, it shouldn’t be about whether ACM is willing to take the risk of a defamation lawsuit; instead, if certain disclosure is indeed a defamation, that means someone’s reputation is, in fact, inappropriately hurt. Putting the legal consideration aside, I am not sure if this is conscientiously right.
Thanks to the committee for preparing this thoughtful article and soliciting community feedback. I am very glad to see ACM move in the direction of more disclosure. The Draft 2 policy is certainly better than the current situation. I also agree that a distinction between minor and major violations is important and the related broad principles outlined are good.
Aligned with comments by Margaret Martonosi and Boris Grot, I do believe, however, that the Draft 2 wording does not go far enough and the language in many places is too vague to inspire confidence that actions taken will be on the side of reducing harm. Clear words are important because otherwise vulnerable members of the community (often the ones subject to harm) will not come forward with complaints.
Let us take the concrete example of the hypothetical case in Sidebar 6. For this case, the only difference between the current and new policies appears to be that the new policy will inform SIG leaders, conference leadership, and CARES (and maybe Frank’s employer) that Frank has been sanctioned. In theory, this is a step forward – the current policy requires leadership to query the database proactively. However, the sad truth of these matters is that Hannah and others in her position are probably junior members and not part of this leadership that will be informed. So while Frank will (hopefully) not be a threat to them in ACM meetings, he will still be around at IEEE and other society conferences, funding agency PI meetings, a senior colleague at a university, interview/give presentations at their universities, etc. When it comes to stopping harm, this sanction unfortunately still leaves many ACM members vulnerable to Frank’s potential bad behavior. If Frank’s research area has conferences in other societies and he is senior enough, a five year ban from ACM will have little impact on him. Moreover, it does not serve as a deterrent to others. And the situation does not provide enough confidence in the system for a junior member to file a complaint. Frank will continue to have a huge influence on Hannah’s career – he will continue to be a reviewer on IEEE papers, etc. (A related topic is ACM’s ability to deal with anonymous complaints were Hannah to complain anonymously or to protect non-anonymous complainants from retaliation.)
Now to the legal liabilities. I do not want to downplay legal exposure for ACM from defamation lawsuits, etc. But if Frank now emboldened by lack of accountability engages in an even more serious sexual assault on Hannah or on her friend in a non-ACM venue, are there any legal implications for ACM? Sadly, since Hannah is likely junior and a victim of assault, the thought of suing ACM is probably not going to be uppermost in her mind. But is this what we are betting on when we weigh the risks and benefits?
The above scenario though hypothetical is inspired by a real analogous situation. We must stop this.
Again, the image I have in my mind of a complainant is a vulnerable and junior member of our community. Filing a complaint is an act of huge bravery today. I don’t think the new draft policy changes that and we must begin there.
Our community is indebted to Huixiang’s friends - as @dchen so eloquently said above, a small number of brave students managed to get attention in this case and change things. Perhaps ACM can consider calling the new policy the Huixiang Chen policy (preferably with edits to make it worthy), that way when people consider applying the policy, they remember why it exists and who it is meant to protect.
My background: I am ACM fellow. As SIGARCH chair, I co-led the formation of CARES. As member of the SIGARCH board as past chair and member of ACM council, I, along with many colleagues in SIGARCH and CARES, spent many hours trying to bring attention to various matters related to investigation, disclosures, and sanctions to ACM, including related to the case of Huixiang. I have first hand account of multiple complaints for violations of ACM policies and have seen how difficult it is for our junior members to file complaints and how current processes can fail them. I am thrilled to see progress (kudos to the committee), but we need more.
In addition to my general comments above, here are some concerns about specific statements:
2:7. 1. Different enforcement policies can limit the information shared by those involved in the complaint, e.g. whether a complainant is allowed to share publicly that the complaint has been filed.
I am not sure of the rationale behind this statement. It seems to me that it should be the complainant’s right to choose whether they share their experience (the subject of the complaint) and their response to the experience (the filing of the complaint) publicly or not. If the complainant abuses this right, the respondent may file a complaint, sue the complainant, or whatever they wish. But it seems to me that it should be the complainant’s right to choose whether they want to subject themselves to this possibility. ACM would not be obligated to protect the complainant and I don’t see how ACM can prevent the complainant from saying what they wish about their experience. This is especially important considering that ACM investigations historically have taken a long time.
Taking the concrete example of the case in Sidebar 6, Hannah may want to describe her experience with or without using Frank’s name so that others don’t go through the same. I am not sure of the rationale for ACM preventing her from doing this.
Also I am not sure what would be classified as “public.” SIGARCH runs a slack channel for women to provide a safe space for various discussions. Would Hannah’s sharing there looking for support violate the policy? Would her just sharing that she had filed a complaint and asking for others who might have gone through a similar experience violate the policy? I can think of so many questions and scenarios where the above statement might unnecessarily limit the support that someone in Hannah’s position may need.
Again, filing a complaint by a junior member is a huge act of bravery these days. We shouldn’t take away the support that such a complainant might need unless there is a clear benefit to doing so.
“should be the complainant’s right to choose whether they share their experience (the subject of the complaint) and their response to the experience (the filing of the complaint) publicly or not”
Totally agree with Sarita’s point on this 
.
(Shan Lu, SIGOPS)
The draft does not say explicitly, but it appears that during an investigation, nobody other than award committees and people directly related to the complaint find out about the investigation. This is a great principle; however, it works well only if the investigations complete in a timely way. Consider a situation where junior people have been bullied to provide favorable paper reviews and they continue to see the bully on program committees and other places where the bully can have influence on their career (because the conference leadership does not know about the investigation) – if this continues for multiple years, it would be hard to make the case to file a complaint. But one needs to balance the rights of the respondent, so the only solution seems to be for ACM to make a commitment to quickly resolve investigations (preferably through the use of professionals).
2.2: When complaints are filed without any evidence or anonymously such that no investigation is reasonably possible, the complaint will be dismissed without investigation…
This comment is not related to disclosure per se as used here, but to the difficulty in filing anonymous complaints to ACM and what it means that “no investigation is reasonably possible.” In my experience, even when a complaint has been accompanied with hard evidence, the investigation was not deemed to be able to be conducted in a manner that would protect the complainant’s identity. This is because often the evidence will reveal the identity of the complainant. We must try harder to ensure whistleblower protection. We must find ways to present evidence that will protect the identity of the complainant (less disclosure) while preserving the rights of the respondent. In my experience, the rights of the respondent have trumped the rights of the complainant often in our processes and we must find a better balance. Again, the basic principle I want to push is that filing a complaint should not be a huge act of courage and bravery, while of course still protecting the rights of the respondent.