AI-generated reviews on Steam are becoming a problem
-
wrote last edited by [email protected]This post did not contain any content.
-
This post did not contain any content.
Review weighting formula needs updates, if it's not taking this into account already. There are many many ways to do this. For example, review and it's score are multiplied by coefficients that are computed from hours spent in the game, percentage of achievements completed, time from the last review posted on the same account, number of people who clicked "this looks like a shopped review" button, etc.
-
Review weighting formula needs updates, if it's not taking this into account already. There are many many ways to do this. For example, review and it's score are multiplied by coefficients that are computed from hours spent in the game, percentage of achievements completed, time from the last review posted on the same account, number of people who clicked "this looks like a shopped review" button, etc.
Sounds good, except which of those metrics you proposed can’t also be gamed by bots?
-
This post did not contain any content.
Kinda sad because I usually rely on steam reviews to see how bad or good a game is.
-
Kinda sad because I usually rely on steam reviews to see how bad or good a game is.
Same. I haven't been let down by it yet, so hearing an AI review surge makes me very sad.
-
Sounds good, except which of those metrics you proposed can’t also be gamed by bots?
Have a lot of those metrics in place & keep the formula private. If leaking the formula into the public seems probable, then make formula polymorphic: certain weights differ based on RNG seeded by hashcode of game's internal ID. This doesn't fully protect from gaming the formula, but it makes automated influence unreliable and hits botters business. It's a questionable approach, but I think it hits botters way more than it hits legitimate reviews, because in legitimate reviews there are zero expectations how certains reviews contribute to overall score. Such expectations can only exist, and thus can only be ruined for malicious actors. This definitely has some limits of how much it can contribute to overall score, because RNG shouldn't be able to make a good game with legitimate reviews not reach good overall score. Unreliable means that botters were able to take 1000$ from client and bump their game, but then they take 1000$ from their next client and their shit suddenly doesn't work anymore for unknown reasons and client is mad and botters decide to quit their business and move to something else.
-
Kinda sad because I usually rely on steam reviews to see how bad or good a game is.
wrote last edited by [email protected]i mean with the review in question its a red flag given the playtime is shown.
steam at least give you some tools to consider if a review is legit. one is the users play time (which is public), another is other reviews theyve made. another is if theyve gotten the key for free.
-
Kinda sad because I usually rely on steam reviews to see how bad or good a game is.
If I'm interested in a game, I usually download a repack to try it out and if I like it, I'll buy it.
-
Kinda sad because I usually rely on steam reviews to see how bad or good a game is.
I've recently discovered your review only counts if you bought it through steam on the steam store.
If you get a key off humble bundle or another site, your review means absolutely nothing. There is a little star next to reviews now that tell you this.
I found it a bit disappointing for steam.
-
I've recently discovered your review only counts if you bought it through steam on the steam store.
If you get a key off humble bundle or another site, your review means absolutely nothing. There is a little star next to reviews now that tell you this.
I found it a bit disappointing for steam.
Umm, so it's harder to manipulate the rating? I mean, seriously, without that requirement, the ratings would be just as worthless as Amazon or Play Store.
-
i mean with the review in question its a red flag given the playtime is shown.
steam at least give you some tools to consider if a review is legit. one is the users play time (which is public), another is other reviews theyve made. another is if theyve gotten the key for free.
The main problem is that those reviews are still included in the overall review score. I think you can filter out low play time but that's an extra step most won't take (because they shouldn't need to)
-
The main problem is that those reviews are still included in the overall review score. I think you can filter out low play time but that's an extra step most won't take (because they shouldn't need to)
wrote last edited by [email protected]would you prefer the platform police it and run into the problem that the google play store/ios app store has about vote manipulation? and purging several reviews (whether legitamate or not)
-
Review weighting formula needs updates, if it's not taking this into account already. There are many many ways to do this. For example, review and it's score are multiplied by coefficients that are computed from hours spent in the game, percentage of achievements completed, time from the last review posted on the same account, number of people who clicked "this looks like a shopped review" button, etc.
wrote last edited by [email protected]Overall reviews: Mostly Positive
Recent Reviews: overwhelmingly negativeReview: this is the worst piece of shit ever made the devs should be hanged!!!
Playtime: 2006 hrs on record
-
would you prefer the platform police it and run into the problem that the google play store/ios app store has about vote manipulation? and purging several reviews (whether legitamate or not)
wrote last edited by [email protected]Steam already does stuff similar to this? Reviews flagged as "suspicious" aren't deleted but aren't included in the overall score instead, and a notice is put next to the overall score when this happens. The same is true for reviews made by people who got the game with a key.
Why not extend this? Like you said, most bots have low play time, valve could exclude (but not delete) reviews with low play time. I agree that doing something like this is a slippery slope towards mobile app store reviews but if it's done right then it is a net positive.
-
Overall reviews: Mostly Positive
Recent Reviews: overwhelmingly negativeReview: this is the worst piece of shit ever made the devs should be hanged!!!
Playtime: 2006 hrs on record
Is this Dark & Darker?
-
This post did not contain any content.
AI slop is a literary and intellectual contaminant the same way PCBs and PFAS are chemical contaminants. And Just like PCBs and PFAS (forever chemicals) it's just as hard to get rid of. But unlike those nasty chemicals we have stupid people insisting it's a good thing and continuing to barf out more of the same AI garbage.
-
If I'm interested in a game, I usually download a repack to try it out and if I like it, I'll buy it.
You could also buy it, try it and refund it if you don’t like it.
-
This post did not contain any content.
The internet is sort of fucked. It was bad enough with marketers ruining search and sites through SEO obsession, but now with this chatbot bullshit everywhere? What's the point? It's all bs.
-
The main problem is that those reviews are still included in the overall review score. I think you can filter out low play time but that's an extra step most won't take (because they shouldn't need to)
People who review games poorly are unlikely to play them for long though. How would you filter them as not to drop most legit negative reviews?
-
This post did not contain any content.
I want my fake reviews to come from a bot farm in some poor country, as god intended.
