Nah. It’s just creepy.
NeatNit@discuss.tchncs.de 4 days ago
I was going to say that this might actually be a good idea, but then I remembered how often AI summaries get key details completely the wrong way around.
I still think it’s a good idea in concept: modding is hard, or so I hear, and this could be a very valuable tool if it worked as it’s supposed to. I doubt it would.
There are some details that I’d have done differently too, if I was somehow forced to implement something like this. For example, for each point in the summary (e.g. “talks about US politics”) there should be a list of comments/posts by the user to exemplify what they say and how they write. It’s useless if the summary doesn’t show you what it’s based on.
Flax_vert@feddit.uk 4 days ago
TheRtRevKaiser@beehaw.org 4 days ago
Lol, you can tell which commenters have never moderated anything in this thread, IMO. If it weren’t for the high likelihood that these summaries will be wrong an appreciable percentage of the time, this would be a huge help for anyone moderating medium traffic subs. Those types of subs, especially if they have relatively hands-on moderation to keep them from being complete cesspools, often involve seeing a comments or post that is borderline, and feeling like you need to go look through the poster’s history to figure out if they’re a bot or a troll. Something like this that actually worked, especially if it linked back to a sampling of the posts/comments that it is referencing, would be a big help in that. Also something like this that summarized a user’s moderation history would be pretty useful.