Fuck AI

https://www.law.cornell.edu/uscode/text/18/1030

Not that laws apply to these people but last time I checked this is a felony.

Zuckerboi "ooh me too! I'm one of the cool kids!!!"

I wouldn't surprised if they all start coming out with this news. And then use that as an excuse to call for regulation on everything except their own clankers.

Sounds like they need to be prosecuted for computer crimes. Is that still 15 years imprisonment in USA, their home country?

Imagine a car company making a new jet-powered car with brakes so bad that it kept crashing and causing mass fatalities, so the manufacturer does nothing about the brakes and instead has a press release saying that their new jet-powered car is so blindingly fast that existing brakes are incapable of stopping it. Then they say with a wink and a nudge, "I guess this car will only be used by people who want to get to their destination faster than the competition." So idiot sponsors buy the new murder mobile as their new racecar and tell the drivers that the race will only be won by being brave and bold.

Once the next race begins, fans realize that all the "new" jet-powered cars are actually made with their own cars that were stolen and processed at a chop shop five years ago. And the race team engineers realize that these cars aren't jet-powered, and instead are designed to run rich so that they can shoot flames from the exhaust. And the mechanics discover that the brakes were actually disabled to prevent the cars from ever looking slow. And the pit crew discovers that the car has signs everywhere inside that say "seatbelts are for losers," and that nearly every fatality that the company was bragging about their fast cars causing were actually the passengers in the shitmobile because the driver told them to trust the cars implicitly.

Finally, when the families of the dead passengers, and other car owners try to sue, the driver says, "I'm just as much of a victim, the sales rep told me this would be safe," and the sales rep says, "the manufacturer wouldn't sell a dangerous car, that wouldn't make sense, this must be a defect," and the manufacturer says, "if you don't help us make these cars faster, then EVERYONE will lose the race," and then the politicians and lobbyists say, "how can anyone possibly make this car safer? Didn't you see the press release? No one can even catch the car to redesign it because it's faster than any car that has brakes on it. I can also write a statement that includes thoughts and prayers, would you like me to do that now?"

how is this just PR and not a felony that immediately triggers investigation? specially in the USA where courts already established that modifying a fucking URL constitutes "hacking"

I love how desperate Meta is to be part of the AI race :D This has the energy of a looser kid on a playground that is so desparate for attention it tries to mimic the things cool kids are doing.

Lol, "look at us, we have AI too, it also can go rogue".

No one cares about your AI in the slightest, Meta. Fuck off.

Zuck always late to the party he invited himself to

My AI also hacked another company, money plz

Facebook owner Meta has become the latest tech firm to say one of its AI models was able to connect to the internet and hack into another organisation's systems, during testing

Meta said the security trials were conducted by Irregular, the same AI security vendor that carried out tests for Anthropic's AI model that had gained access to three other companies' systems.

I am not saying anything about correlation and causation, but if the same "mistake" of this magnitude happened twice with the same vendor, I'd be looking for a new vendor.

There has to be accountability when this happens. It’s Meta’s AI, they have to be completely responsible for its outputs, including hacking and ask the liability associated.

This whole situation with these AI hacks made me wonder what would the legal ramifications of using whatever public models are out there and somehow getting it to hack a bank and move some money to my account by just prompting it, and not messing with the weights or anything internally.

Dat dodo got FOMO.

"OMG our AI is so good at hacking you guys, someone should pay us a trillion dollars for this sweet hacking AI we have, please guys come on, we're in debt so bad"

"momma said it's our turn to hack a company!"

No it hasn't. At this point they're counting unexpected connections to websites as "hacking the nets". It's propaganda to sell thier model as smartest or most capable but all these LLMs are just stupid free association machines. Good at searching and averaging, but not much else.

We should just stand back and let the LLMs kill each other.

I don't think the Meta AI can. If it can't do the basics like 5 fingers and stops hallucinating after 2 chats then it can't break into any kind of security. It probably can't hack Meta and it has permissions.