Stack Overflow volunteer moderators down tools over secret new policy that obstructs removal of AI-generated content

Posted on Sunday, Jun 18, 2023 by Chris Hayner

Featured in this episode of Chaos Lever

I don’t normally just copy and paste the headline from the article I’m linking, but in this case I felt compelled to. “Down Tools” is an English phrase that means ‘go on strike.’ So with that in mind the headline doesn’t sound crazy.

So there- you learned something today. You’re welcome.

ANYWAY- In the war against AI generated content wades people working for one of the sites whose content was prime pickings for training AI- Stack Overflow. The venerable site used by many a developer who wanted to pretend to know what they were doing has instituted a curious instruction to its moderators. Do not remove AI-generated content “outside of exceedingly narrow circumstances.”

There’s a part of this where the company is concerned that AI detectors are not reliable, and content might not truly be AI-generated, which- fair. There’s another part where the moderators who are striking are saying there’s more to the policy that they’re not allowed to publicly share, which- mysterious.

I assume that the company is all too happy to allow AI to post because it will increase their magical ‘engagement’ numbers- This is the same reason why Twitter is fine with bots in reality, because as we all know, 80% of twitter is bots.

Twitter, though, is concerned with shitposting. StackOverflow, though, is concerned with correct answers. And as we all know, AI can’t be trusted for that, no matter how much engagement it offers.