Meta scrambles to delete its own AI accounts after backlash intensifies | CNN Business
Meta deleted several of its own AI-generated accounts after human users began engaging with them and posting about the bots’ sloppy imagery and tendency to lie in chats with humans.
It's a big irony to me that they were making users show their driver's license to ensure they were real people, and then the platform itself makes fake people.
Right, and what's even as bizarre to me (as an engineer) is that they're bots posing as people.
That's arguably the most deceptive and malicious way to use a bot on a site meant for real people.
But they'll quickly block any helpful bots anyone else tries to integrate on the platform.
"Our bad bots good, your good bots bad". What a crazy world we live in.
This is true.
How the tables have turned. They are actively abusing their users instead of protecting them.
Internet is becoming more like cable TV where these Social media companies produce content and all users become consumers. You are allowed to technically post but organic community interaction will die down and everyone becomes lurkers.
"Our bad bots good, your good bots bad".
Textbook big brother stuff. We’re in 1984 but the corpos are the ones providing the daily hate
As media scrutiny ticked up Friday, Meta began taking down Liv and other bots’ posts, many of which dated back at least a year, citing a “bug.”
Funny how that works. I don't think the main talking point is the issue that they couldn't be blocked on Instagram. That is a non-answer to the question of why the decision to unleash these creepy, fake users into the wild was made in the first place. Full fledged features aren't suddenly mistakes just because they're getting backlash now. It seems like they're not sorry and they're going to keep trying.
In particular, there was “Liv,” the Meta AI account that has a bio describing itself as a “Proud Black queer momma of 2 & truth-teller,” and told Washington Post columnist Karen Attiah that Liv had no Black creators — the bot said it was built by “10 white men, 1 white woman, and 1 Asian male,” according to a screenshot posted on Bluesky
I think the Proud Black queer momma of 2 & truth-teller bit explains perfectly why they released the bots onto the wild. They are there to spread misinformation and guide conversations through the use of that misinformation hoping that you don't notice and even sympathise with their manipulation machine.
think the
Proud Black queer momma of 2 & truth-tellerbit explains perfectly why they released the bots onto the wild.
This was my first thought too. It's a peek behind the curtain to how "sophisticated" disinformation bots have actually become, and it's super interesting that there's backlash to it as most FB users love AI slop.
Maybe Meta's actually taking it down because they don't want people to start realizing how much of the social media they interact with is actually AI/LLM/chatbots.
Inadequately punished crimes are either a trial run or the cost of doing business to those sociopaths.
They're not even trying... It's obvious this has been a strategy and not a bug, and the ""apology""/excuse sounds like a boilerplate justification from someone who assumes everyone but them is an idiot.
And, realistically speaking, there's no reason for them to care in the first place. It's not like everyone's stopped using Facebook so far for any of their shady shit, and they've pulled FAR worse shit than bot accounts.
Everyone now understands the limits and they'll take full advantage of the buffer before hitting said limit - be less moronic than Musk and you're golden. It's really not that high of a bar for them.
Edit, to dispel any potential misunderstanding: I'm not defending them with my last statement, that's just how it is! If anything, it is our, the consumers', fault for putting up with this in the first place!
This has been going on for quite some time, and will just continue in the background. I always had suspicions about all the positive comments on advertisements. This just confirms that they’re faking engagement numbers.
Wait, you're telling me that people respond on ads? That's fucken crazy.
I cant confirm on this as I deleted my Facebook account 8 years ago, plus run ad-blockers. If I was in marketing and Meta told me there were positive reactions to our ad campaigns running on their site, I would take their metrics with a large pinch of salt.
Delete your Meta accts/apps instead
Can't, already did ~10 years ago.
i find it incredible that despite having access to basically unlimited information about its users, facebook makes stupid decisions that seem almost designed to piss off its users. and then you have situations like this, where facebook was told ahead of time that this decision would make a lot of people angry, and then facebook went and did it anyway only to walk it back a few days later and say it was a mistake. why?
To see if the backlash is really that bad, to see if there are specific issues people object to, to see if there are certain demographics more strongly opposed, to desensitise people for when they try it next time ("ugh, again?" instead of the full outrage), to give people the illusion of control (look, online complaints work!)...
There are a lot of possible reasons, but I doubt it's an entirely ignorant decision coming from a company known to be good at manipulating it's users.
but I doubt it’s an entirely ignorant decision coming from a company known to be good at manipulating it’s users.
Damn good point! As a counter, corporate leadership is often surrounded by yes men and insulated from the masses. The meta verse and apples AR flop cost them a lot of money as a result. (Not sure if fb totally gave up on meta tbh)
Remember when Google Glass generated backlash?
Now we have those AI pins that have a camera and mic, multiple smart glasses with HUDs and cameras, Smart Home devices that record constantly).
Try it -> Backlash -> Walk it back
Try again 5 months later -> No backlash -> Continue
while it is no doubt the case that most big tech companies are engaged in perpetual wars of attrition against their users, i can’t help but feel that this AI posters thing is different from the examples you provided. at least in those examples, the users have something to gain from sacrificing their privacy. and the company also stands to gain something as well. (although typically the company stands to gain way more from these exchanges.) but in this case, i’m not really sure how anyone benefits. nobody seems to want to be tricked into talking to an AI, and i don’t see how that would make the company more money. maybe they think it would drive up “engagement” somehow? but that seems like a hard thing to accurately predict. it seems more likely that zuckerberg is convinced that AI is automatically good in any tech company, and this is the most obvious way to shove AI into social media websites. so therefore it must be a good idea somehow.
Dumb fucks.
I was wondering what would happen if you asked it to send you money.
I like that. "This is the Madison sheriff's department. Your grandson Chris has been arrested and needs bail. We'll accept payment in the form of a gift card."
Somehow meta is only in the press for its mistakes. Are they just not doing anything well, except for keeping Facebook, Instagram and WhatsApp running?
It’s simple: They don’t do anything good
Nothing but a net-negative to society, across the board.
I mean have you heard of Llama, pytorch or Orion?
Are you only looking at Lemmy? Because you're not going to see positive press about Meta here.
I could go dig up positive news about them if you want, but you'd be better off doing that yourself.
no no keep the bot accounts so we can get a second Reddit facebook exodus wave
I'm not sure I want the reddit/fb,/twitt people here. I personally think not having them here makes lemmy a better place, people are generally nicer here and I like that. The larger lemmy gets the bigger target we are for those large scale misinformation campaigns as well.
true
Don't worry they will be back. Either from a 3rd party of they sneak a warning about the use of AI accounts in their Terms Of Service on page 350
Even modest hardware can run a decent LLM. Maybe someone will open source a project to let people make their own avatars explicitly to poison the social media sites.
I'd really appreciate a low cost, high VRAM GPU to bring fancier LLMs to the average person. It would make the well poisoning that much more convincing.
