Well, well, well, what do we have here? The guardians of Stack Overflow, those volunteer moderators who’ve turned the site into their personal fiefdom, are having a tantrum. As of June 5th, 2023, they’ve gone on a historic general moderation strike, joined by over 850 contributors and users.
Their beef? Stack Overflow, Inc. isn’t giving them the attention they feel they deserve regarding being able to police AI-generated answers. To which I say, “Welcome to the club, mates.”
These self-appointed stewards of quality, who’ve spent years closing valid questions and deleting good answers on a whim, are suddenly concerned about the core mission of the Stack Exchange network. The irony is almost too delicious to swallow.
One of their primary grievances revolves around a new policy restricting AI-generated content removal. They fear it will lead to an influx of inaccurate information and plagiarism, eroding trust in Stack Overflow. I can’t help but chuckle because trust was eroded years ago, as far as I am concerned.
Then they whine about recent policies that allegedly undermine the autonomy of individual Stack Exchange communities, leaving them feeling marginalised and disempowered. Sounds familiar, doesn’t it? Thanks to their heavy-handed moderation, it’s the same feeling countless new users have experienced.
And I also want to point out that while I am generalising here, not all moderators are responsible for the demise of Stack Overflow. But, if you search Hacker News, blogs and other parts of the web, you’ll find plenty of tales of heavy-handed moderation. You can even see it for yourself by browsing through Stack Overflow.
In their strike, they’ve decided to suspend activities like flagging, closing posts, and more. They demand a policy change to allow them to enforce policies against AI-generated answers. Meanwhile, Stack Overflow, Inc. is holding its ground, pointing out that the tools moderators used to detect AI-generated content had a high rate of false positives, leading to unnecessary suspensions of innocent users.
We’ve seen the same situation play out in academia. Students are being falsy accused of using ChatGPT and other AI tools, even though their claimed accuracy is far from the truth. So, I am not siding with the moderators because there is currently no definitive way to know if something is AI-generated. Sometimes there are tells, but oftentimes, you can’t tell.
An even more bonkers story is about a Texas professor that failed an entire class because his detection methods said the entire class used ChatGPT. Basically, in this instance, the professor just asked ChatGPT if it wrote the texts he fed it, and it said yes for everyone. The chances of every student cheating would be incredibly low. This is another example of heavy-handed and accusatory accusations around AI usage.
So, to the striking moderators, I say this: You’re not the victims here. You’re part of the problem. You’ve spent years running Stack Overflow like a private club, alienating new users and stifling the spirit of the community. Don’t act surprised when the tables turn. You’ve been ignoring users for years.