Researchers secretly experimented on Reddit users with AI-generated comments
Researchers secretly experimented on Reddit users with AI-generated comments
www.engadget.com
Researchers secretly experimented on Reddit users with AI-generated comments
28 comments
-
Saying this like reddit hasn't been bot infested for a decade.
I mean there were some genuine looking bots well before LLMs and AI, and even then you could just be lazy and make a copy post bot that would repost old content for karma farming while the terminally online userbaae would upvote your slop for you.
-
The bots would pretend to be black in order to say “as a black man, I don’t like BLM” or pretend to be a male rape victim so they could say “the experience wasn’t as traumatic as some would say”. How the fuck do they reconcile this as ethical? They’re actively arguing with real people and acting like it’s all just random data. Social scientists once again de-humanizing their subjects, I guess
-
The experiment is bad, they've not controlled for anything. Their bots are most likely arguing with other bots.
-
-
Secret? As opposed to all the blatant AI bot accounts?
-
Read the article instead of responding to the title. It was a university conducting formal research, which created AI bots that impersonated different identities. "As a black man..." style posts in ChangeMyView.
The subreddit mods issued a formal complaint to the university when they learned of it, but the university is choosing not to block its publishing on the grounds of lack of harm.
-
Yeah how do they know they were interacting with real users and not another researcher/troll bot?
-
Ummm. They knew, guys.
Also this is extremely violating. And this bullshit,
. “We believe the potential benefits of this research substantially outweigh its risks"
What kind of psychotic sssholes run an experiment on an unsuspecting public because they believed that it's ok to fraudulently engage with others without theirspotential subjects being aware of it
This is something trump would do
-
It's Switzerland. Morality there is measured in $ (the same goes for their "neutrality")
-
-
-
So they figured it was a good idea to use a racially-charged fake profile to provoke online users for an ‘experiment’? And one would assume these responses were subsequently studied without the poster’s consent?
That’s going to run afoul of a few European privacy rules I imagine. Someone definitely needs to get fired, blackballed and sued for this. At the very least.
-
My assumption is that a majority of the content we see on all "social" platforms these days is AI.
-
We are all AI on this blessed day
-
-
Wow. There's no way this was actually greenlit by an ethics committee unless they're all corrupt. This is so blatantly wrong and manipulative.
-
It was reviewed by an ethics board, but they changed the parameters of their study (shifting more towards using personal info against commenters/making up personal info to use in arguments) without it being approved by the ethics board. Essentially their study wasn’t going anywhere so they made it more unethical in order to have more data to work with. They’re on Reddit trying to defend themselves but no one is having it
-
There are no ethics in the game of human life. The world is far worse than this.
-
Just so you know, this is called Moral Nihilism, and it seems to me about as useful as a physicist saying, "none of our models are 100% accurate, therefore they are all completely useless".
Like, yeah we agree that ethics is a construct we've invented, and no ethical system is perfect, and none of them ever will be. But that doesn't translate to them all being useless. I know we're living in a stressful time, and I understand the feeling of wanting to just flip the table and give up, but please recognize that as a purely emotional response, not a rational one.
-
-
-
Reddit sold user content so it could be used for AI training! If they're mad or take legal action, it's only because they didn't get a piece of the action...
-
Well, it USED to be a good place to ask a question because there were so many people on it, there was likely to be several experts on _______ at any given time. Then it got more and more left, then more and more right, and after a while it felt like being on twitter again. What I'm saying is that their value and name became pretty worthless, this just makes it 0.
28 comments