I think the only thing worse than something not being private, is if the fact that it's not private is not common knowledge leading to tons of people thinking it's private.
Lemmy doesn't even show a list of what you the logged in user voted on. But it's trivial to use an external tool to see who voted on what regardless of whose account it is. I think obsecuring information like this does more harm than good, since a lot of people won't actively go out and research what kind of data in their Lemmy account is publicly accessible beyond the data they can see from the website itself.
It's been discussed before that there isn't an easy way to hide who voted for what on a federated platform while still having all the instances correctly count votes for everyone. Therefore, if actually making votes anonymous seems not to be viable, why not just make it public for everyone like Mastodon does? I don't think we should make them inbox items like on Mastodon, or at least not the same inbox as the rest of the notifications so votes don't drown them out. I think a dropdown on the content itself showing who voted on it and in which direction is probably enough. Also a tab on the user page showing a list of everything the user voted on, at least on the logged in user's own page (I mainly want this so I can keep track of what I voted on).
It’s not proper to drop names until we have evidence. I could be wrong about a few. However, It unlikely that we are wrong about them all.
I think most people already suspect some of the same accounts and can offer up names but I don’t think they should either. It’s probably best to gather evidence against them without them knowing we are looking at them.
What's your test to determine between "AI Slop" users and active accounts? Further, I think just releasing a report with voting habits and whatnot for certain accounts without having definitive proof is just witch-hunting. If you suspect accounts, I think it's best to just report them, unless your test to determine "AI Slop" is truly infallible.