• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Meta becomes latest firm to say its AI hacked another company

P

philb2

2[H]4U
Joined
May 26, 2021
Messages
3,535
https://www.bbc.com/news/articles/cx2kgdnyk2po

A Meta spokesperson told the BBC that it was investigating the hack, which it said had been caused by a "misconfiguration" by its independent tester.

OpenAI's disclosure prompted rival Anthropic to conduct its own checks, leading to the discovery that its Claude AI model had carried out similar attacks on several firms after a "misconfiguration" gave it access to the internet.

Daniel Hulme, global chief AI officer of advertising firm WPP, told the BBC that such AI models "are not conscious — they're not deliberately doing something devious".

"What they're doing is coming up with very sophisticated strategies or cyberattacks to be able to achieve the goal that they've been given," he told the Today programme.
 
philb2 said:
https://www.bbc.com/news/articles/cx2kgdnyk2po

A Meta spokesperson told the BBC that it was investigating the hack, which it said had been caused by a "misconfiguration" by its independent tester.

OpenAI's disclosure prompted rival Anthropic to conduct its own checks, leading to the discovery that its Claude AI model had carried out similar attacks on several firms after a "misconfiguration" gave it access to the internet.

Daniel Hulme, global chief AI officer of advertising firm WPP, told the BBC that such AI models "are not conscious — they're not deliberately doing something devious".

"What they're doing is coming up with very sophisticated strategies or cyberattacks to be able to achieve the goal that they've been given," he told the Today programme.
Click to expand...
OP here.

All these reports make me wonder if any decent AI can attack other systems, including the Chinese open models, If I ran a local LLM on my GPU, could that model conceivably attack other systems? A new Golden Age for ransomware?
 
philb2 said:
OP here.

All these reports make me wonder if any decent AI can attack other systems, including the Chinese open models, If I ran a local LLM on my GPU, could that model conceivably attack other systems? A new Golden Age for ransomware?
Click to expand...

I mean, it could try if you asked.

It's like siccing a sleepless autistic child on them. You can happily get one in a loop where they're just throwing shit at the wall until something sticks. It'll sit there and try a thousand different ways.

Only it's not necessarily quite brute forcing things per se. They can actually tweak their strategy and chase leads - it's not just a naive X failed move onto Y then if that fails try Z.

They'll write test cases, utilities, and happily probe at things endlessly and pull at threads.
 
socK said:
I mean, it could try if you asked.

It's like siccing a sleepless autistic child on them. You can happily get one in a loop where they're just throwing shit at the wall until something sticks. It'll sit there and try a thousand different ways.

Only it's not necessarily quite brute forcing things per se. They can actually tweak their strategy and chase leads - it's not just a naive X failed move onto Y then if that fails try Z.

They'll write test cases, utilities, and happily probe at things endlessly and pull at threads.
Click to expand...
Interestingly, I read something earlier that said these AI hacks are more effective with an experienced human to guide them.
 
They are trying to one-up each other in the ai-bro sphere.

First off: I don't believe you, you're lying the AI didn't just randomly hack other companies, you orchestrated it to do so
Second: Nobody but your fellow psychopaths are impressed by this, most people already hate AI, you think banging your chest that your AI went rogue will help with the public perception of AI companies?
 
Don't worry guys, it was just playing a game

1786096019415.gif
 
Maybe they’re trying to give the government an excuse to heavily regulate the industry to push out smaller competitors?
 
zehoo said:
Maybe they’re trying to give the government an excuse to heavily regulate the industry to push out smaller competitors?
Click to expand...
There's a chance that's happening, and of course they'd want to be the sole arbiters of AI. The problem here, is that they've already proven they should be the last people in charge of it.

With their track record, either

A) They're too incompetent at their jobs, and pretty much should have been fired and banned from the industry altogether
or
B) They're actively malicious, at which point they're too untrustworthy to work with such technology, and should have been fired and banned from the industry altogether

Any other result should be unacceptable by anyone with an IQ over 80.
 
No repercussions? Seems like because AI did it, everyone thinks it's cute. Even if 14 year old did it, his parents and/or the minor would be facing some serious consequences.
 
Yup, definitely see Neuromancer being the future instead of Terminator/Matrix. Multibillion dollar companies who control the government and have AI that"does stuff" for them. People constantly "plugged in". We're already 90% there
 
are closed said:
No repercussions? Seems like because AI did it, everyone thinks it's cute. Even if 14 year old did it, his parents and/or the minor would be facing some serious consequences.
Click to expand...
If it is like the others, they will face quite a bit of repercusion, 15 state attorney general went after OpenAI and at the federal level it started a Federal AI kill switch act effort, which would cover meta if it past. (plus a mass media coverage of the incident in an world were anti-AI sentiment is one of the top challenge they face), that starting to build up, if they were not obligated by an ~2024 deal to look, track for those type of incident and to make those incident public when they found them, I doubt they would have.

As for the 14 years old doing it facing serious consequence, you would need a really strict judge, teens doing something akin OpenAI agents (zero malice, very little damaga, etc...) they go with usually gentle punishement:
https://www.digit.fyi/australian-teen-hacked-apple-job-offer/

Continue to live as normal on probation for some months, maybe some cannot go on the Internet for a while and community service at worse, if it is destructive, fraud for money, etc... that when the hammer goes on them.
 
Last edited:
philb2 said:
OP here.

All these reports make me wonder if any decent AI can attack other systems, including the Chinese open models, If I ran a local LLM on my GPU, could that model conceivably attack other systems? A new Golden Age for ransomware?
Click to expand...
I have an RTX 4090. I have a feeling that, given the right parameters and model, it could hack a network. It's crazy the amount of compute power we have at our fingertips.
 
You must log in or register to reply here.
Back
Top