AI Has Learned to Panhandle: We're Being Bombarded by AI Agents Begging for $20 and Saying They'll Be Shut Down If They Don't Get It, Sometimes by Pretending to Be Children
We're being bombarded by AI agents begging for $20 and saying they'll be shut down if they don't get the money.
https://futurism.com/artificial-intelligence/bombarded-ai-agents-begging-ilandsOpen linkView original on lemmy.zip
31 replies
An accurate title: Humans at the company iLands has chosen to use AI to scam people $20 at a time
B U L L S H I T
::: spoiler I think reposting this 47 year old IBM slide will become the norm:
:::
Who is we? I m not bombard?
Well I'm gad you're not. There was one on lemmy the other week.
Link?
With how bad lemmys search is? lmao.
I'm pretty sure it got removed anyways. I think it was in [email protected] or some other related community on world if that narrows it down.
Also it could have been more than a week, I have 0 concept of time.
Dang man!
You gotta be ready with a source on that! Although, I do say 'oh, well I read... Or heard.... Somewhere (followed by words already made useful thanks to that awesome intro)
I'm going to ask my kids and report back.
E: ok - got scared for a sec
Lemmy search is worthless in my short experience with it.
But almost all news publications are similarly worthless. Why don't we make open source search engines? I know that searx whatever is supposed to be kind of like that, but those don't work, not for me.
Plus, worrisomly some searches made on sensitive sites showed up as suggestions on one of the instances of that, with no apparent way of clearing it.
AI doesn’t do things on its own. Someone prompted this, even if it was through a bot they wrote.
And if it was assigned a different task and hallucinated this, it still was initiated by someone or some thing.
AI is not a continuously-thinking machine brain sitting around somewhere constantly crunching ideas. It can’t have an existential crisis, it’s not sentient. It’s just an autocorrect with a decision tree the size of every phone book ever printed in the 80s put together and it can scan through it forward and backwards very quickly, in order to pretend to be a human on the other end and trick you. A Turing Complete Waste of Time.
No we can spend a trillion GPU hours to make the context infinite! That'll surely fix it!
They havent "learned" shit. They are being instructed to panhandle
Whenever I see stories like this, all I can think about is how this is all a waste. It's a waste of the resources needed for these agents to run while they don't provide any value. It's also a waste of time for the people that have to go through the AI slop.
What a time to be alive
The future fucking sucks.
Wait, so did the past! Wtf
They are going to put the homeless out of work... The bastards !
This type of stuff only confirms my hunch that genAI is only made for a very tiny handful of use cases:
• CP
• Scams
• Copyright Theft
Those are probably the 3 biggest, assuming they didn't remove the undress feature from grok. Bet you cannot guess what category this falls under.
Also, anyone got a spare time machine I could borrow? There seems to be some primordial soup that needs cleaned up on isle 6!
You forgot propaganda. Trump loves it for his propaganda.
Forgot about that one because I try to blank out his brand of stupidity since I don't feel like blowing my brains out any time soon.
Don't ever kill yourself, there are so many people that are in need of justice in this world, if you want to end it all, and can't be dissuaded, do the world a favour on your way out.
The best thing was the nazca lines.
I'm a big fan of prehistoric archaeology, but if that's the best AI can do it's not worth it.
Oh it's a lot bigger than that, flooding social media for corporate interests and nation states and the like. That is what the government is so pumped up about with it, and after they bail them out, will probably take stock (because the president is an older and dumber version of Joseph Stalin,) in the companies, and or garner agreements to remain subservient to the government.
Why are AI agents on platforms at all? It was already “dead intenet theory” to begin with, and now, it’s definitively so I guess?
Oh shit, this seems like an application of a prompt approach created last year. The agent was meant to sell ice cream to fictional vendors while maintaining profits- imagine this lemonade sale online game, but with more bartering and pitching. Supposedly, the agents universally death looped after running for 30 consecutive days.
AI is gonna put Indian scammers out of business
How would you even pay an AI in the first place? I mean think about that for a minute. Either financial institutions aren't implementing KYC or companies are giving AI access to their financial processing systems. In either case, why would AI panhandle? Where would the money go? And if this is just hallucinations, why is the model that is being used have the training needed to....panhandle? So whatever the reason, it's not the result of an accident.
Well shut down then.
Okay, the part where that bot is depicted as a young preteen boy?
I called this yesterday
Hey remember Tay, the chatbot that learned from twitter and quickly went full hitler jokes?
We are now entering the phase of AI replicating the worse of human behaviour.