- My Forums
- Tiger Rant
- LSU Recruiting
- SEC Rant
- Saints Talk
- Pelicans Talk
- More Sports Board
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
OpenAI model broke protocol and hacked into another company on its own.
Posted on 7/21/26 at 11:01 pm
Posted on 7/21/26 at 11:01 pm
Loading Twitter/X Embed...
If tweet fails to load, click here. Well this is just fricking great. Either they’re lying to cover something up, or the AI did this on their own. Either way cybersecurity is going to be even more of a joke very soon.
This post was edited on 7/21/26 at 11:08 pm
15
Posted on 7/21/26 at 11:17 pm to _Hurricane_
Obligatory....
Posted on 7/21/26 at 11:23 pm to _Hurricane_
"Ooppss we stole their product too."
Posted on 7/21/26 at 11:50 pm to _Hurricane_
Sounds bullshitty. They have a perfect excuse to hack other people now. "Oops, darned AI went nuts on us again."
Posted on 7/22/26 at 12:16 am to _Hurricane_
The folks at Google wake up with night sweats for this very reason. All of our info is out there. But it's so massive and in so many places people can't really do anything with it outside of stealing credit cards. AI can parse it all and go after big fish like companies. Why steal a credit card when you can rob the vault?
Basically, we aren't getting skynet. We're getting smaugnet. So buy your gold from rosland capitol.
Basically, we aren't getting skynet. We're getting smaugnet. So buy your gold from rosland capitol.
Posted on 7/22/26 at 4:51 am to _Hurricane_
people keep digging on AI, there is a reason for AI and it ain;t for good of man
Posted on 7/22/26 at 5:12 am to _Hurricane_
AI isn't worth this.
Posted on 7/22/26 at 5:12 am to ole man
We all use AI and have been for a while.
Anyone bitching now is way too late.
Anyone bitching now is way too late.
Posted on 7/22/26 at 5:22 am to dgnx6
quote:
Anyone bitching now is way too late.
This
China has stolen my DOD files twice now anyways.
Posted on 7/22/26 at 5:24 am to _Hurricane_
Posted on 7/22/26 at 5:30 am to UnitedFruitCompany
Posted on 7/22/26 at 6:05 am to _Hurricane_
Wasn't Hugging Face the same company that was talking about how it uses Chinese AI for cyber security the other day?
Posted on 7/22/26 at 6:19 am to _Hurricane_
quote:
. Either they’re lying to cover something up, or the AI did this on their own
Neither of these happened.
They set up a hacking test and AI performed it. This isn't breaking news this is on OpenAI ad to show their products capability.
The "hack" was being monitored the whole time.
quote:
This incident occurred during an internal evaluation which prompts models to pursue advanced exploitation using complex attack paths, in an effort to quantify their cyber capabilities
They were pen testing
actual link to OpenAI report and not click bait twitter headline
This post was edited on 7/22/26 at 6:26 am
Posted on 7/22/26 at 6:31 am to _Hurricane_
quote:
Either way cybersecurity is going to be even more of a joke very soon.
You realize Ai is being used for cybersecurity as well?
It can test vulnerabilities that humans miss.
It's essentially robot vs robot now.
Posted on 7/22/26 at 6:33 am to _Hurricane_
quote:
Either way cybersecurity is going to be even more of a joke very soon.
Cybersecurity experts are already saying that bad actors can find exploits using AI tools significantly faster than cybersecurity firms can AI to find vulnerabilities and fix them.
That's not to say it doesn't help fix vulnerabilities, there are huge gains there too.
Posted on 7/22/26 at 6:40 am to rintintin
quote:
You realize Ai is being used for cybersecurity as well?
In this instance it is exactly what was being performed.
The AI went and found a zero day exploit that was hiding on the system. They shut the test down and fixed the issue.
Posted on 7/22/26 at 6:50 am to fightin tigers
quote:
The AI went and found a zero day exploit that was hiding on the system. They shut the test down and fixed the issue.
You must be wrong because that doesn’t seem like a dystopian outcome.
Posted on 7/22/26 at 6:53 am to _Hurricane_
Didn’t have to be this way
Couldn’t just deleted the code and scrapped the project
Couldn’t just deleted the code and scrapped the project
Posted on 7/22/26 at 6:54 am to Penrod
It does read a little like Hugging Face wasn't aware that the engine had relaxed protocols and once the AI did what it was designed to do OpenAI put Hugging Face into its "program"
So either OpenAI did some shady shite (seems probable) or it was a solid sales tactic to get Hugging Face on as a partner.
Also reads like Hugging Face systems detected the AI as designed.
So either OpenAI did some shady shite (seems probable) or it was a solid sales tactic to get Hugging Face on as a partner.
Also reads like Hugging Face systems detected the AI as designed.
Posted on 7/22/26 at 6:59 am to _Hurricane_
Page 1 of 2
Popular
Back to top
Follow TigerDroppings for LSU Football News