RE: Report as Steem Representative, 04.05.2024
On the one hand, I'm glad that you had the same experience as me. On the other hand, it's really particularly sad in both cases.
It did actually cause me a bit of conflict... on one hand (I've got to copy your expression here), we don't want people pretending to be somebody else, and using AI to generate content. On the other hand, the content appeared to be authentic and well written - creating a better impression for external people visiting the site, and presumably search engines. Which made me wonder, should it be allowed? (Of course, the reward element complicates things further.)
I'm pretty sure I've debated this, or similar with @remlaps before where AI could be used to attract new users. It's the "cheating" element that bothers me.
Yeah, I don't think the AI aspect really matters much. If you use AI and acknowledge that you're using AI, fine and dandy. If you write fiction and acknowledge that you're writing fiction, fine and dandy. But...
If you use AI and claim (imply) that it's human or you write fiction and claim that it's fact, then it's a problem. I can't see this author's posts any more, so I have no opinion on the immediate topic.
I felt similarly conflicted with a couple authors recently who (I strongly suspect) are using some tool that I can't identify to generate content that's not obviously abusive and not obviously created by AI, so I was stymied about how to curate those. Again, it had the feeling that someone was cheating. Although I couldn't say for sure, and I wasn't sure what the harm was, anyway.
I ultimately decided to give low-percentage upvotes, and it seems that the votes were low enough that they moved on to other communities. Which, I guess, makes them someone else's problem, at least for now. I'm not going to level public accusations if I can't say what the cheating mechanism is. :-(
Of course, detecting these sorts of deceptions at scale is another matter, entirely. It occurs to me that we could fund a "fraud department" out of the SPS, but that just devolves into the problems of Quis custodiet ipsos custodes and how to measure success. For example, I've seen no evidence that STEEM WATCHER has produced a meaningful reduction, and I have no reason to believe that the "detectives" aren't just creating spam in order to collect rewards for reporting it. In the end, I think we need automation or else SBD interest at a high enough rate to disincentivize cheating for posting rewards.
It's difficult, and becoming increasingly so. There are often other telling signs, beyond the content, that can help to validate a suspicion. In this instance, there were 3 other factors which alone (even from an individual account perspective) could have been dismissed but combined and across multiple accounts, became more than just coincidence.
I think that this is what happens on Hive and they take a huge payment every day to do so. Perhaps they have more success because they have the power to deal with the biggest farmers that use upvu or tipu and the likes of trafalgar - thereby (theoretically) improving their Trending page. Their Trending page is definitely more attractive than ours. Whether it's value for money or not... It's $290 per day.
The comments underneath their proposal is eye opening though.
This won't be true of all Steem Watchers, but it's certainly my opinion on some / many of them. (Disclaimer: I don't look at any of their reports so can't comment on whether the work that they do is genuine or beneficial.)
This relates to a recent Netflix video that we watched about child exploitation. I said to my family that the reason we find these sorts of problems in the theme parks and schools and children's entertainment industry is simple. Predators go where the prey is.
As you note, it's probably the same concept with money & power.
On Steem Watchers, I also haven't read many of their reports, so my impression is just superficial, too. It just seems to me that if there's an easy way to game the system, then someone's going to make use of it. And that looks like an easy way to game the system.
One thing they’ve done well is to limit the number of reports to 1 per person per week so the rewards will get spread out.
True. That's a point that I hadn't really considered.