Effective immediately, the moderators of Academia.SE (wrzlprmft, cag51, and Bryan Krause) are on strike. This is part of the network-wide action described here, and follows an Academia moderator resignation a few days ago. We will not perform any moderation functions until this situation is resolved. Other users are welcome to join us by refusing to perform moderation functions such as voting, voting to close and editing posts (but everyone will make their own choice; please don't harass users who continue to participate normally).
Why is this happening?
We moderators have been ordered not to intervene when AI-generated content is posted to our site, except under very limited circumstances. While the exact details are not public, even the publicly-available guidance effectively allows almost all AI-generated content site-wide, and admits that "this standard would exclude most suspensions issued to date." In other words: our policy banning AI-generated answers is unenforceable; automatically-generated content cannot be treated differently than human-generated content.
It is hardly necessary to explain why this is a problem generally (and others already have). But here on Academia.SE, we are particularly concerned because many of our users are asking for advice on life-altering career decisions. In such a setting, automatically-generated content is not a mere nuisance, but can cause irreparable damage to someone's life. While we recognize that AI-generated content is increasingly difficult to detect, and there will be false positives and false negatives, the current policy (don't do anything at all) is the worst possible response.
Is this an overreaction?
Not really, this seems to be our only option short of allowing AI-generated content to run rampant over our network. Moderators network-wide have been pushing back against this policy for a week already to no avail. On the contrary, we've been told about another policy in the pipeline that will make the situation even worse. And the company has historically shown little flexibility even when they were clearly wrong**; as a result company-moderator trust is not very high.
When will this end?
Our sole demand is that any network-wide policies must allow this stack to delete the majority of automatically-generated content and message/suspend users who post it. We acknowledge that there will be false positives and false negatives, and so we are willing to take guidance. But a policy that effectively allows all automatically-generated content is unacceptable.
Alternatively, this will end when we all get fired and new moderators are elected. That's fine, we don't get paid (even the swag we were promised never arrived), and there is little point in sticking around so we can impotently preside over a hellscape where bots talk to bots. In that case....we appreciate you for having elected us; so long, and thanks for all the fish.
** For context: in 2019, a popular moderator was fired and slandered publicly; after a huge uproar, the company issued a lawyer-speak apology but refused to reinstate the moderator. Instead, their solution was to form a "moderator council" to make us feel heard, but the council never had much influence and has since disbanded. Even so, the individual community managers have been great, and we had successfully built a certain amount of trust before this incident destroyed everything again.