Phrasing it in the form of a question, since I really don't know the answer, but there's a pretty big issue going on right now with users (at least 20 so far that I've seen) defacing and/or deleting their content in protest of the recent OpenAI announcement. This is happening on multiple sites that I've found.

End users can only see this when the post can't be deleted because either:

  • There's an accepted answer
  • The user has deleted too many posts already

Without linking to the post, there's one on Ask Ubuntu that was defaced with the following text:

I have removed this answer in protest of Steal Overflow's decision to collude with OpenAI.

However, once we identify a user as taking this action, it's then possible to search SEDE for other content they have recently deleted. In this case, the user has also deleted posts on Stack Overflow and Super User (at least). Since I have 10k on those sites, I can see and flag those for undeletion and rollback.

However, the user also has accounts on several other sites that I won't be able to flag.

In this case, they've also updated their profile page to claim that all of their content is AI-generated. This is not the case. They have simply done this in order to try to force deletion through another means.

Do Mods have a good way to look for this behavior? Some users are simply going to delete their posts in protest without also defacing the answer. How will a Mod be able to differentiate between a legitimate deletion vs. one done in protest (i.e., a defacing deletion)?

  • 12
    Some turns up at Charcoal HQ, particularly the ones with gibberish-edits or obscenities. We then rollback or flag for it to be done.
    – W.O.
    Commented May 8 at 15:36
  • 9
    Can't answer for mods on all sites but in mod private places I hang out there is awareness and mod use their tools to address the issues that arise.
    – rene
    Commented May 8 at 15:46
  • 3
    "Do Mods have a good way to look for this behavior?" As a user, revert the edit if you can. If they do it again, mod flag for vandalism. That usually does it. If enough reversions happen, an automatic flag is raised. A mod flag where you explain what's happening would be the first way to go, IMO. Speaking as a mod.
    – Mast
    Commented May 8 at 15:52
  • 3
    Are all mods aware this is happening? No. I can guarantee you some are not. Luckily, in most situations it's not necessary for all of them to be aware if the mod flags are used correctly. Poking one in your local chat room could help depending on the site/mod as well.
    – Mast
    Commented May 8 at 15:53
  • 12
    If sufficient content is self-deleted, an automatic flag will be raised as well. There are dashboards we can look at too. Raising a flag on content of theirs that wasn't deleted can work too. If that all fails and you've tried to contact a local mod in chat and that failed too, I know some mods will take a mod flag on their content to be notified. Make sure to add plenty of context in those cases or people get confused. If that still fails and there's a major problem, find a friendly mod from a different site and ask them for a recommendation.
    – Mast
    Commented May 8 at 15:59
  • 10
    If the problem is big enough, mods talk to each other a lot about a lot of different things. We have our tricks. And if that fails too, there's always the option of contacting the team. Defacing content is a license violation and if none of the mods are willing to act on that, there are bigger problems that could warrant poking someone about.
    – Mast
    Commented May 8 at 16:00
  • 4
    @Luna I'll "squint" and read "license" in this case as "Terms of Service", which it pretty much is. A ToS and licensed-use-of-the-site could be used interchangeably. I truly don't think Mast mean the CC license in that sentence. Commented May 8 at 16:59
  • 3
  • 5
    While not strictly "made aware", there's also a limit of 5 deletions per day, and 10k "moderator tools" to check recently deleted posts. Commented May 8 at 17:56
  • 8
    @MetaAndrewT. And unfortunately, while most never notice, the 10k tools do not show self-deleted content, only content that was deleted by Mods, Roomba'd, or Community vote Commented May 8 at 18:03
  • 3
    While it's true that as a user without the ability to see deleted posts it's difficult to raise flags on deleted posts, primarily because the system doesn't show you the post, it is possible for any user with the ability to raise flags on the site to raise an "in need of moderator intervention" flag on deleted posts. It can be done either through the SE API or through SE's internal flagging endpoint. That doesn't, really, make much of a difference, as it's quite rare for users to know that it's possible or to have a convenient way to raise such flags.
    – Makyen
    Commented May 8 at 20:06
  • 3
    Reaching the deletion limit of 5 also triggers an autoflag. There's only so much we can check though, and obvious vandalism is easier to handle (because searching for "openai" is enough for many of the edits, and the destructive rest get caught by smokey). Deletions are tricker, but not impossible to manage. Not sure how lower-volume sites handle it though
    – Zoe
    Commented May 8 at 23:07
  • 3
    You could, of course, just also take note of the fact that pretty much the entire community is extremely displeased with the openAI announcement and make it clear to staff that they really need to rethink their stance posthaste before more people leave. Commented May 9 at 8:23
  • 3
  • 2
    Maybe the vandalism is the minor problem here, since it is merely the final symptom of the user permanently leaving the network never to come back. Now that is the actual problem. And at the point where they are going berserk on their own content, it's too late to fix, they are already gone.
    – Lundin
    Commented May 15 at 6:44

2 Answers 2


I would strongly encourage folks not to spend a lot of time and effort trying to detect this stuff. Flag what is obvious or obnoxious.

It is impossible for normal users to vandalize a post such that it cannot be recovered; the system has multiple checks in place to restrict widespread abuse.

The system also has rate limits that will thwart efforts by normal users to effectively monitor such vandalism, especially when it is subtle.

Employees, OTOH, have tools with much more freedom to monitor such things. Given this is all a reaction to various money-making schemes by the company, let them spend some of that money on handling the fallout.

  • 7
    First time I think I've ever downvoted Shog9 ;-). I understand (and yet ... don't) the anger about OpenAI. I've been fighting bad AI answers here probably more than anyone else, and I really don't think I'm exaggerating when I say that. But I actually see the OpenAI/SE deal as a win-win, as much as it can be. The content was already out there, scraped, used for training, etc. SE could go the "we'll sue" route like so many others (and I'm sure there was the threat of that) ... Commented May 9 at 14:51
  • 12
    ... but the deal that they brokered with both Google and OpenAI appears to require that responses from those LLMs will be attributed, and link back to the content here. That's an improvement over the existing situation. And I hope we all assume that SE is trying to make money on our answers - Hasn't that been the whole business model from the start? (not a rhetorical question - You worked there, so I'm hoping you have some insight) Commented May 9 at 14:52
  • 18
    I'm hopeful too, @NotTheDr01ds - but hope doesn't moderate vandalism, hope doesn't write code, hope doesn't enforce the terms of business arrangements, and hope certainly doesn't stand in for clear and honest communication about those terms. We're gonna need to see a lot more before we know what the real fallout is gonna be. As a former employee, I know what I'd have been doing right now in this situation, and I expect nothing less from current staff.
    – Shog9
    Commented May 9 at 14:57
  • 3
    @NotTheDr01ds Why should unpaid volunteers keep mitigating the bad decision making of people who profit from their work already? Let Mr. Chandrasekar et al. reap what they've sown.
    – Dan Mašek
    Commented May 9 at 15:01
  • 9
    @DanMašek cause to a certain extent the company hasn't been great stewards for years - I don't think the goal is to enable the company to do whatever, but rather ensure the damage to our communities is minimised in spite of those decisions, and ensuring protests have a positive impact that's more than burning down your account Commented May 9 at 15:09
  • I doubt the company will spend the money/effort on network sites besides SO and perhaps a couple of the other bigguns, on the rest of the network it's going to be up to moderators and users, realistically. (I trust CMs will look into things if it comes to their attention, I just don't believe that they'll be assigned sufficient priority towards hunting for it) Commented May 9 at 15:17
  • 3
    Well, the cost needn't be huge @Bryan - a few minutes running a query a periodically would suffice. Vs. normal users hammering SEDE once a week and then filtering out false-positives by hand because SEDE doesn't have revision history for deleted posts.
    – Shog9
    Commented May 9 at 16:09
  • @Shog9 I didn't say the cost needed to be huge, only that I don't expect it to be spent. Commented May 9 at 16:36
  • 4
    I was not very fond of the idea that AI will train on our content for quite some time now (it has been over a year since the initial discussion about selling content for here on Meta) and company could definitely do a better job when announcing things and communication with users here. However, our content has been used for training AI a long before all that so all anyone can do is not post new content if the don't want to be source for AI training. Vandalizing the content that is already part of AI only hurts people that don't want to use AI as a source of information. It is my duty to clean. Commented May 9 at 18:23
  • 4
    I am not serving the company here, I am serving our community. Commented May 9 at 18:26
  • 4
    @JourneymanGeek I'm well aware of that (corp not being good stewards). And I certainly don't endorse any vandalism, I frown on that (it's against the spirit of the site, against the license agreement, rather childish, etc.). But let's be honest, the relationship between the company and the community has reached exploitative levels long ago. It's unhealthy, and there's a point where you need to say "that's enough". Sunk-cost fallacy... I've been through that a few times already.
    – Dan Mašek
    Commented May 9 at 18:33
  • 7
    @ResistanceIsFutile I'm not saying "don't clean". I'm saying, this is the company's mess to clean up. We've got folks in the comments talking about using SEDE to identify this stuff, using 10k tools, etc. - all of those suck for mass deletion. There are mod tools for handling mass deletion, but if those aren't sufficient then the solution isn't "burn insane amounts of time working around deficiencies in the system". This is literally a "few minutes" type of job for someone with employee access, so either they're gonna clean up their own mess, or you should seriously question things.
    – Shog9
    Commented May 9 at 19:16
  • 2
    Every once in a while I get reminded that letting you go was really bad day for the whole network. This is one of those days. I am afraid that if we leave this to become company's problem it will never get resolved. Commented May 9 at 19:28
  • 6
    I'm not worried, @Resistance - this is hardly the first time these sites have faced mass vandalism, and years ago my team devoted a lot of time and effort into developing and tuning systems to minimize the damage. Assuming those systems are still in place, still tuned, this is at most an annoyance. Folks are upset, and they have some reason to be - while I don't think this is an effective form of protest, I get it. And frankly, if it helps some folks' peace of mind to do a bit of vandalism and move on, so be it - as long as we don't let it waste too much time or effort.
    – Shog9
    Commented May 9 at 19:49
  • 1
    I am not worried either. It is not a nice situation, for sure, but we still have way bigger problems with users that are actively posting AI generated answers. Commented May 9 at 21:31

To a significant extent, we're going to have to rely on eyeball MK I - on any Stack Exchange site smaller than Stack Overflow, a deletion is still going to bump a post, and defacements are generally obvious.

I also suspect once the initial shock/anger wears off, instances of this should taper off.

Other than potential volume, and a good many folks here seem to be users with old accounts, I hope this is a wake up call of how proper, clear messaging to the community is as important as press releases. While we can handle the vandalism here to a point - there's a load of bad press, and hot takes floating around on tech media that's likely to amplify the problem.

  • 1
    "a good many folks here seem to be users with old accounts" - Yes, but in some unscientific observation, it also seems that most of those users haven't posted anything new in years (even after checking for deleted content). Commented May 9 at 14:43
  • 1
    Pretty much the very unscientific conclusion I made, Oh and trolls Commented May 9 at 14:45
  • 3
    Post deletions don't bump the question, though...
    – Catija
    Commented May 9 at 15:19
  • Ah oops. Coulda sworn it did Commented May 9 at 15:30

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .