Wikipedia talk:Responding to threats of harm

For similar proposals in the past, see: Wikipedia:Responding to suicidal individuals, Wikipedia:Helping suicidal individuals, and Wikipedia:Threats of violence.

Samaritans & co

WP:SUICIDE and similar redirect here; there should be a hatnote linking to agencies that support the suicidal, such as the Samaritans. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:44, 18 November 2021 (UTC)

I moved the existing link to nearer the top [1] EEng 19:16, 18 November 2021 (UTC)

Thank you. I missed that, so I'm sure someone in crisis would have done so. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:29, 18 November 2021 (UTC)

emergency@wikimedia.org connections are refused

Last night I encountered an IP user who threatened two other editors. One was threatened "...I will kill you..." while the other was threatened with being stabbed in the chest. An email to emergency@wikimedia.org was immediately returned as "Connection refused due to abuse". (The threatening editor and his/her sock has since been blocked and the offending editorial summaries redacted.)

With the emergency@wikimedia.org email account essentially shut down how should threats of physical violance be handled? Blue Riband► 21:01, 26 April 2022 (UTC)

Can you email cawikimedia.org to ask about this? (You can think of that email address as the "non-emergency number" for essentially the same team.) Mz7 (talk) 06:56, 27 April 2022 (UTC)

@Mz7 - I took your advice and emailed ca@wikimedia.org but that also immeditely bounced back "due to abuse":

The following addresses had permanent fatal errors
- ca@wikimedia.org (reason: 550 5.7.1 <ca@wikimedia.org>... H:MB[wikimedia.org]Connection refused due to abuse) (expanded from: ca@wikimedia.org)

Transcript of session follows ----- ... while talking to localhost: >>> RCPT To:<ca@wikimedia.org> <<< 550 5.7.1 <ca@wikimedia.org>... H:MB[wikimedia.org]
Connection refused due to abuse 550 5.1.1 ca@wikimedia.org... User unknown Blue Riband► 15:12, 27 April 2022 (UTC)

My initial impression is that this is likely to be an issue between your email provider and the WMF servers (or somewhere upstream). Possibly your provider (not you personally, but someone using or spoofing the same provider) has been used to send a large amount of spam and so has been added to a list. If this is the case then the best thing to do is to raise the issue with your provider (giving them the full unredacted email headers) so they can get themselves removed from the relevant list. If you have another email account you could also try making ca@ aware of the issue, again giving them the full email headers - it might be possible for them to do something. Thryduulf (talk) 22:44, 27 April 2022 (UTC)

@Thryduulf - that appears to be the case. When I ran the address for my email host through mxtoolbox it came up on a Blacklist UCEPROTECTL3. Blue Riband► 04:20, 28 April 2022 (UTC)

Another bounce-back problem

Data point: today someone again reported email to emergency@ bouncing back. [2] EEng 01:19, 2 August 2022 (UTC)

Could we not have an alternative means, such as a mobile number used only to receive text messages, something of that sort? I guess various social media alternatives are possible with suitable privacy settings. Chiswick Chap (talk) 08:58, 2 August 2022 (UTC)

Texts are a bad idea for many reasons. These two recent incidents are the first ever, from what we know. My thinking is we might just add a note to the instructions to the effect that if your email bounces back you move on to the "Notify admins" step, mentioning that your email bounced back. Something simple like that. EEng 15:50, 2 August 2022 (UTC)

Would email an arbcom or steward be a viable step? Slywriter (talk) 16:29, 2 August 2022 (UTC)

I wonder if the best option is to suggest if it bounces back to email oversight via the contact form? AFAIK the oversight list tends to be monitored fairly well even at off hours in the US and Europe as it's covered by several people compared to the risk of trying to contact a single editor. (That includes both that the editor will be unavailable or won't see the message, but also that it might be transparently hidden or rejected. At least with the WMF they seem to be sending bounce back emails so people know their contact attempted failed even if they probably shouldn't be doing that.) Also since you're using the contact form, I wonder if the risk of it bouncing is lower although there might be some risk a bounce will be suppressed. Yes, editors won't be able to add images or other attachments, but the idea is for this to be a first step in cases where stuff has broken down. Probably should explain to the people on the oversight what we plan to do and why first though in case they have objections although I assume they've already got some experience with people contacting them with stuff which they feel needs to be passed on to the emergency team, and the emergency team is also used to contacting them when stuff needs suppression. Nil Einne (talk) 04:24, 3 August 2022 (UTC)

I really think we're overreacting. We have no idea whether this is going to be some kind of chronic problem, or is just some blip. So far it's a blip. EEng 04:36, 3 August 2022 (UTC)

I just realised one thing is the contact form can't be used by editors without an account or without an address on their account or who aren't willing to share this address. There is the oversight-en-wp email address but if it has the same filters and sensitivity won't be useful. Also the form thing is perhaps a moot point since there is Special:EmailUser/Emergency as well. So on further consideration, not sure my suggestion is very helpful. Nil Einne (talk) 09:38, 3 August 2022 (UTC)

I'm not sure what you mean about filters for the OS email, but anyone can email us, and unless it gets blocked by the spam filter (which given how much spam gets through, is pretty unlikely) it will show up. Primefac (talk) 17:31, 4 August 2022 (UTC)

Plan

I think that we need to have a proper contingency plan in place for this. I do not know if there have been any threats to commit suicide on this site; since this article exists, I would presume so. The fact of the matter is that suicide threats can, of course, be distressing. Some editors might be making their first ever suggestion in a talk page, when they see that a new message has come in two minutes ago of somebody threatening to commit suicide. Can you imagine how distressing that would be? Now imagine this on a high traffic page, maybe even the main page. All hell would break loose. My idea is that we have a group working alongside the emergency response team, made up of mainly admins and other trusted/often online individuals. It could also be possible, although difficult, to have a bot (a webcrawler, if you like) that searches talk pages for certain keywords: suicide, kill you, kill myself, etc. When these are detected, it is deleted and the revision is deleted, it is filtered through a group of people (or more precisely, whoever is online in that group) who can decide if it is plausible and/or specific. They can then work out the person's general area and contact local authorities if it is one or both, or discuss on their talk page or in a more private way. The other option is that there is no bot, but an post detailing the threat and the url is sent to a talk page containing people in the group mentioned above. They can then take necessary action, described above. I hope these contingency plans/ideas help you. Crystalpalace6810 (talk) 16:52, 14 June 2022 (UTC)

This is pretty much how we operate at the moment. Primefac (talk) 17:29, 14 June 2022 (UTC)

Please elaborate, CrystalPalace6810 described two options, and I want to use my own recent experience (see my block log) as a case study so i can better understand what to do next time. NewsAndEventsGuy (talk) 01:27, 31 July 2022 (UTC)

Right, and both options are pretty much how we operate. The VRTS OS queue receives tons of emails per day, some of which involve this subject matter. There is also #wikipedia-en-revdel ^connect, where we get plenty of requests daily (less often for this type of material, but it does happen). The third option, which generally is more on the "dealing with the individual" side of things, is to email T&S directly; they don't tend to deal with the on-wiki stuff themselves but will often let an OSer know there's something that needs hiding. Primefac (talk) 15:23, 31 July 2022 (UTC)

Thanks Primefac, I'll work on aborbing all this.... and thanks for the link to VRTS which I never heard of before (probably because it just didn't register) NewsAndEventsGuy (talk) 15:42, 31 July 2022 (UTC)

@Primefac.... Never heard of T&S... I want check so I read the right things... Please confirm you mean Trust and Safety? NewsAndEventsGuy (talk) 14:15, 1 August 2022 (UTC)

Yes. Primefac (talk) 14:19, 1 August 2022 (UTC)

Suicidal ideation via talk page comments that stop short of explicit imminent threat of *physical* harm

The project page seems to instruct us to report explicit threats of imminent physical harm. What should we do when an editor reveals that they simply feel depressed enough to consider suicide? NewsAndEventsGuy (talk) 01:31, 31 July 2022 (UTC)

Tamzin your input here would be welcome NewsAndEventsGuy (talk) 01:32, 31 July 2022 (UTC)

The word "imminent" is not used at all on this page - any and all threats of self-harm should be dealt with in the manner described on the page. If you are indeed taking your example statement seriously (step #1), then you should contact T&S (step #2) for the user and an admin (step #3) for the content. Primefac (talk) 15:26, 31 July 2022 (UTC)

I'm a bit confused because at my usertalk [3], admin Tamzin explained that I'm afraid no, we don't have a good comprehensive set of options for dealing with suicidal editors. For outright suicide threats (not "I've been suicidal lately" but "I'm going to do it tonight"), the WMF takes jurisdiction under WP:EMERGENCY, and local oversighters support by suppressing the threats themselves (both to avoid suicide contagion and because a suicidal person can't be said to have meaningfully consented to saying that.) But for things that aren't really an imminent threat, it's more an ad hoc mixture of administrative actions. @Primefac (or anyone else too), do you agree that "I've been suicidal lately" or things "aren't really an immient threat" do not t qualify for reporting to WP:EMERGENCY? The more certain we are that we are on the same clearly-articulated page, the better we'll be at helping editors in need, and that's the reason I'm trying to get my brain around this topic.NewsAndEventsGuy (talk) 15:56, 31 July 2022 (UTC)

I believe that your quote and what you are asking are not as directly connected as you think. What Tamzin is saying is that we have no one-size-fits-all answer to every solution; the clear-cut cases are clear-cut but everything outwith that ends up being down to discretion and circumstance. However, the same steps should always be taken; the outcome, particularly whether to remove/RD/OS/etc, is really the only thing that will change.

What you are asking (or rather, asking me to agree with) is that the imminence of a threat determines what course of action we take, and I would argue that it doesn't (all but the aforementioned "last step" of whether to remove/RD/OS something of course); Tamzin correctly points out that even someone saying content like this can be distressing and cause a cascade effect for other editors who might be triggered by such statements, and so we need to approach each situation with the care and diligence it requires. We are volunteers on this project, and the instructions provided here are provided so that we do not have to deal with any more than strictly necessary; in other words, our main "job" in these situations is to let someone who knows how to deal with this stuff handle it, or at the very least pass it up the chain so that it can get there. Primefac (talk) 16:44, 31 July 2022 (UTC)

In a nutshell, I arrived here confused about three things. We have resolved two.

1. "imminent" (resolved, that's not a factor)

2. "explicit threat of physical harm" (resolved, EMERGENCY is not limited to such statements)

3. take all threats seriously.... I am still confused. If an editor says "I'm really hating myself these days", "Having a really hard time, wonder what's the point" or "I hate my fucking life" or "even thought about suicide" or other vague expressions of "being really down", is the foundation and/or community hoping any editor seeing such remarks to implement WP:EMERGENCY? Or do those sorts of comments fall in a grey "judgment call" category? Or something else?

Thanks NewsAndEventsGuy (talk) 17:24, 31 July 2022 (UTC)

Or do those sorts of comments fall in a grey "judgment call" category - yes, in short. I say this often when people request oversight, but it's better for the team to decline to act on something than not know about something important in the first place. If you are in doubt, ask someone (discreetly). Primefac (talk) 17:46, 31 July 2022 (UTC)

(edit conflict × 2) So it's not a matter of whether it's an imminent threat: It's a matter of whether it's a threat at all. "I've been suicidal lately" is not a threat, even if it's deeply concerning. In this case, the remark in question was called in to -en-revdel. I made the initial assessment that no admin action was needed, but that monitoring was; no one challenged that assessment. I watched all of GoodDay's edits for the subsequent 24 hours or so, until I was satisfied the statement wasn't going to morph into a threat. I stand by that decision, because, again, "threat" is the operative word here. Once we start taking administrative action against vaguer statements, where do we draw the line? What do we do about someone with a "This user sometimes struggles with suicidal ideation" userbox? Never mind that this page also covers non-suicidal self-injury, something that between 10% and 30% of young Westerners will engage in at some point in their lives; do we take action against "sometimes I self-harm"?

I agree that a better-defined course of action for non-threatening-but-still-troubling references to suicide (and NSSI, and violence against others) would be beneficial. The revdel/OS side is primarily about protecting people from the consequences of things said while not able to consent to those consequences, but I don't think that's the issue here. Perhaps a template to email someone saying, essentially, "I hope you're okay, and please let me know if I can help [in the form of directing you to appropriate resources because I am a volunteer administrator and not a mental health professional, or, if I am the latter, not your mental health professional], but also, going forward, could you please not discuss things so bluntly on-wiki? It may negatively affect other editors."

I will shill briefly here for User:Tamzin/Guidance for editors with mental illnesses, and also note User:Tamzin/Discretionary admin things § Self-requested blocks, which specifically concerns mental health. Also, N&EG, I saw your email. I hope to find time to respond today. -- Tamzin^{[cetacean needed]} (she|they|xe) 17:49, 31 July 2022 (UTC)

Thanks Tamzin, no hurry on the email followup. For that matter I only check the account every couple of days. I'd rather use myself as the test case or invent a fictious one. Like in fire fighting, you sometimes think the fire's out but it flares up again. My apologies to User:GoodDay, I didn't mean for your situation to be further aired here, as I try to get up to speed with how Wiki handles these things. NewsAndEventsGuy (talk) 17:57, 31 July 2022 (UTC)

No problem & I feel no shame. GoodDay (talk) 17:59, 31 July 2022 (UTC)

I'm 'no longer' suicidal. But, I will try & help here, anyway I can concerning mental health. GoodDay (talk) 16:56, 31 July 2022 (UTC)

May you never have to go through it again! NewsAndEventsGuy (talk) 17:02, 31 July 2022 (UTC)

Fresh start to kick around ideas

A worthy topic has emerged from all this.... could we formulate any better ways handle "non-threatening" cases?

For starters, I'd like to restate what I think I have learned above, so if I miss the mark here someone please correct me. As I understand the status quo, we conceptualize these cases as falling into one of two sub-groups

• Where editors retain Agency (psychology), i.e., the mental capacity to make reasoned decisions, and
• (More seriously) when they don't.

As I understand it, we think of the first group as still having sufficient presence of mind to give "consent", and their comments about suicide are not considered a "threat" as that word relates to the protocols in Wikipedia:Responding to threats of harm. In contrast, the inability to give consent (e.g., during times of incapacitating mental or emotional anguish) lowers the threshold for what sort of comments will be considered a "threat" and thereby open the door to intervention by authorized community members.

I'm going to pause now, to invite a checkin...... I've tried to re-state my education on this stuff. How am I doing so far? NewsAndEventsGuy (talk) 14:41, 1 August 2022 (UTC)

I kinda have an idea concerning how to 'lower' the stress of a Wikipedia editor, aside from walking away from the project. As we know, sometimes content disputes can get heated. But, I think such 'heat' can be avoided, if one or more editors involved in a content dispute, stuck to the topic at hand. We all must avoid personal attacks in content disputes, of course. A kinder, gentler discussion, is the best environment for all editors. GoodDay (talk) 01:09, 2 August 2022 (UTC)

Isn't that already covered by WP:NPA ? NewsAndEventsGuy (talk) 02:36, 2 August 2022 (UTC)

So it is. Gosh, I am getting old. GoodDay (talk) 02:46, 2 August 2022 (UTC)

Heh, I thought it was just me with that problem. NewsAndEventsGuy (talk) 09:45, 2 August 2022 (UTC)

• needs discussion This is the kind of issue where we need a highly organized, global and multilingual, well documented public discussion. Wikimedia Foundation staff have made presentations about this at Wiki conferences, but those generally are not well documented and also the discussion takes a different tone when it is staff organized versus community led. The response that I would want is a community led planned discussion with parts on various language Wikipedias and other parts in coordinated public video chats recorded and published on wiki and in accessible platforms like YouTube.

In Wiki LGBT+ we are dealing with some other big issues which are beyond what volunteers can manage, and beyond what can be resolved with in-person conference meetings. Part of the response we proposed to address this is requesting Wikimedia Foundation sponsorship to organize discussions which take everyone's comment and also summarize the results. I have a grant proposal here -

• meta:Wikimedia LGBT+/conversation series proposal

I think the same kind of response could work for addressing suicide. Some barriers to this include 1) we have no identified wiki community members available for hire to take money to organize discussions 2) it is unprecedented and we lack social structure to manage more professional conversations like this 3) there are some social barriers between what Wikimedia Foundation does and what the Wikimedia community does, and we need more discussion on who sets values and ethics and how and 4) conversations like this are bigger than Wikipedia and we need better planning for presenting any practices we develop to the public.

In the case of both LGBT+ issues and suicide response, if the Wikipedia community were to take positions, then probably journalists would report that and other communities would copy us. We have some really powerful advantages that we have an engaged community who will discuss such things and also that we are stakeholders in these issues who actually have great experience with these challenges.

If anyone can think of a way to turn money and sponsorship into planning and written best practices, then I support them in apply for grants. I do not want to do this myself, but I would support others in doing this. Bluerasberry (talk) 13:19, 2 August 2022 (UTC)

Thanks so much Bluerasberry. If I don't comment substantively soon, this was great and I'm already thinking about it. Good luck with the LGBT+ issues you are trying to address. A young half clone of mine thanks you and everyone trying to healthily address that topic. NewsAndEventsGuy (talk) 14:26, 2 August 2022 (UTC)

@Blueraspberry:, I've read through your comment a few days and I'm still uncertain that it is you think needs discussion. Are you referring to the existing process described on the project page attached tot this talk page (shortcut WP:EMERGENCY)? Or to the idea of building on that for instances of editor's admitting suicidal thoughts/feelings of a less "emergency" nature? NewsAndEventsGuy (talk) 00:42, 5 August 2022 (UTC)

Comment. Understanding human nature a bit, I don't want to see Wikipedia becoming a psychiatric monitoring service that helps locking up, drugging, or treating people against their will. Thinker78 (talk) 20:35, 4 August 2022 (UTC)

WP:COMPULSORY applies. PRAXIDICAE🌈 21:00, 4 August 2022 (UTC)

This is useless

ARen't we supposed to contact the emergency email if a user seems to have intent to commit suicide? Cause I just did so and guess what? Absolutely nothing happened. I had emailed the Trust and Safety team as well as the emergency email regarding a user who seemed to want to commit suicide right then and there. And yet nothing ever happened, and now it appears that they have gone ahead and done so. This angers me greatly, what the hell is the point of this even existing if they're going to do nothing about it. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 16:25, 8 November 2022 (UTC)

You make the assumption that nothing was or has been done, but you quite literally know nothing about the situation other than a couple of talk page posts. You assume something has happened, but you do not know.

You are supposed to contact T&S, but T&S is not empowered with magical fairy dust that allows them to stop someone doing something. Hell, at best they can contact local law enforcement, and even then it's hit or miss whether anything can be done. Primefac (talk) 18:49, 8 November 2022 (UTC)

And how the hell would you know? You're not someone who has seen multiple people kill themself while just sitting helplessly. No one does. No one ever will. YEs i Know that they can't guarantee they will stop someone, however I Figured they would at least try to talk to the user, rather than just let everything fricking happen. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 18:52, 8 November 2022 (UTC)

I'm on ArbCom. I'm an Oversighter. ArbCom had a phone call with the WMF last night. Should I continue? Primefac (talk) 18:55, 8 November 2022 (UTC)

I don't see what ArbCom having a phone call with the WMF last night has anything to do with this at all. ― Blaze Wolf^Talk_{Blaze Wolf#6545} 19:00, 8 November 2022 (UTC)

ArbCom has a monthly call with the WMF to discuss issues related to the English Wikipedia, including issues that cannot be otherwise discussed in an open forum. If you will not read between the lines on this, please at least check your email. Primefac (talk) 19:08, 8 November 2022 (UTC)

Sadly, it is unclear if law enforcement or mental health professionals do more good than harm to people who seem to have intent to commit suicide. A trip to a mental health hospital can destroy the life of someone who could have just been joking or written something without actually intending to do it. And mental health "treatment" can actually violate fundamental human rights in many occassions and countries.

Wikipedia should extraordinarily limit what it involves itself in law enforcement cases. Or else it could end up having in-house cops to monitor talk pages or being governments' informant. Thinker78 (talk) 03:24, 11 November 2022 (UTC)

• Agree, inadequate I know nothing about this particular case but I advocate for general reform because I have community involvement in other general cases.

We the wiki community of editors and the general public require public, de-identified information about crisis processes. Neither the Wikimedia Foundation nor anyone else has ethical or moral standing to claim exclusive management of our crisis responses or for judging the extent to which they are correct and sufficient. Here are some starting requests:

1. How many suicide reports do we get in a year
2. What data do we have about the outcomes
3. Where is the response process documented, when will a rationale be published for why it needs to be private, and what third party has critiqued its appropriateness
4. Who decided the resource allocation for suicide response, and when can this investment be discussed in public with the wiki community
5. Where is community participation in this process? Community members bear the burden of outcomes, but are not obviously represented in the Wikimedia Foundation's design and management of the process.

At meta:Talk:Private_Incident_Reporting_System#Get_data_to_third-party_researchers I request de-identified data for incidents to go to third party researchers. That applies to suicide also. Among other stakeholder groups for this data, meta:Wikimedia LGBT+ wants representation in this general discussion. Bluerasberry (talk) 19:58, 8 November 2022 (UTC)

Bluerasberry, you're going to have a lot more luck asking the WMF directly. Posting it here is unlikely to be seen by them, much less to get a response. GeneralNotability (talk) 22:38, 8 November 2022 (UTC)

The WMF prefers to get requests in the form of grant proposals and I have one at meta:Grants:Programs/Wikimedia Community Fund/Conversation series - LGBT+. If they give the money then I will coordinate with Wikimedia LGBT+ to hire someone to ask the question to WMF, the community, and the world, and publish the result.

These kinds of questions take ~100 hours to ask and 1000 hours for a first answer. I am not sure that I can get an answer but I want to get someone to scope the issue.

As an aside, ArbCom needs a paid secretary from the Wikimedia community because it is a waste of labor to elect 10 super skilled Wikimedia volunteers, have them meet with WMF staff who each consume US$200,000+/year, then not publish public reports of the outcomes. I appreciate ArbCom a lot and things used to be run by volunteers but somewhere along the way the WMF started bringing in $200 million / year and the investment in ArbCom needs to be increased from $0 to more than $0. That money needs to go to community because WMF lacks standing of community stakeholders. Bluerasberry (talk) 22:52, 8 November 2022 (UTC)

@Blaze Wolf: Here is a blog post from 2014 giving a little bit of detail of what happens behind the scenes. FWIW I've sent maybe 5-10 emails to emergency@ and always received a "we'll look into it" response within a few minutes. If you didn't get that, there may have been a technical glitch, and one that should be looked into before it happens again. But you won't necessarily get anything more. Suffusion of Yellow (talk) 21:16, 9 November 2022 (UTC)