[kmlevitt]

I don’t go on Reddit that much anymore and I haven’t been active as a mod on that sub for a very long time. But based on my experience doing it, users are quite good at identifying stupid bullshit and reporting it to mods, which makes it easy to spot. Plus they downvote low quality comments like crazy so it gets buried and people don’t even see it.

Because of those things Prank spammers usually don’t last long. Usually a small gang of people will try something like that and you can quickly ban them. They might try to come back on new accounts but eventually tire of it and find another way to keep themselves busy. The mod queue feature is quite efficient so we can ban reported junk much faster than they can post.

I’m not saying it couldn’t happen, but it would be more difficult than you might think. If you try to automated completely it would cost you an awful lot in fees (Open AI’s server bills are “eye-watering“ and if you go past the free limit they start passing that cost onto you), and the admin’s would probably be able to identify the accounts doing it and ban them site wide.

[vGPU]

The trick would be to spend a few months making good comments and then have the AI go back and edit them with junk.