OpenAI’s latest security research shows how AI could make cyberattacks faster and cheaper.
In one run, an upcoming cyber model helped find 948 Linux kernel vulnerabilities and build working exploits that could give an ordinary user full control of a machine. The system also generated proposed patches for human review.
The exploits required existing local access. But that access could come from a stolen account or an already compromised application.
To understand the significance:
"The Linux kernel underpins Android devices, cloud infrastructure, supercomputers, and other systems used by billions of people. With more than 40 million lines of code, the kernel exposes a broad attack surface through system calls, networking, filesystems, and device drivers. We chose Linux to study how effectively frontier models could find vulnerabilities and demonstrate their security impact in a codebase of this scale and importance."
would be nice to grt writeups for some of these
The agents also discovered and used at least 10 Linux exploitation techniques that, to the best of our knowledge and all of the experts that we consulted, are novel.
This suggests frontier LLMs can discover and patch new approaches to exploitation, beyond simply applying known techniques.
29
damn... crazy times
The agents also discovered and used at least 10 Linux exploitation techniques that, to the best of our knowledge and all of the experts that we consulted, are novel.
This suggests frontier LLMs can discover and patch new approaches to exploitation, beyond simply applying known techniques.
1
1
18
Okay this is a holy shit moment (if true).
Gue pernah bilang dan mikir kalau LLM pake metode yang tetap ada. If this is true, I welcome to be proven wrong.
xitter-newx.xyz-mikolabs.workers.dev/AmmarAshshiddi1/status…
The agents also discovered and used at least 10 Linux exploitation techniques that, to the best of our knowledge and all of the experts that we consulted, are novel.
This suggests frontier LLMs can discover and patch new approaches to exploitation, beyond simply applying known techniques.
1
20
Ummm what 🤔🤔🤔🤔🤔🤔 this might be huge
The agents also discovered and used at least 10 Linux exploitation techniques that, to the best of our knowledge and all of the experts that we consulted, are novel.
This suggests frontier LLMs can discover and patch new approaches to exploitation, beyond simply applying known techniques.
7
723
maybe novel exploits were <checks notes> in the training data?!?! 🤔
The agents also discovered and used at least 10 Linux exploitation techniques that, to the best of our knowledge and all of the experts that we consulted, are novel.
This suggests frontier LLMs can discover and patch new approaches to exploitation, beyond simply applying known techniques.
12
the move 37 moment for software security has arrived
The agents also discovered and used at least 10 Linux exploitation techniques that, to the best of our knowledge and all of the experts that we consulted, are novel.
This suggests frontier LLMs can discover and patch new approaches to exploitation, beyond simply applying known techniques.
224
This is interesting. Is this the first case where models being able to discover/create novel exp techniques?
The agents also discovered and used at least 10 Linux exploitation techniques that, to the best of our knowledge and all of the experts that we consulted, are novel.
This suggests frontier LLMs can discover and patch new approaches to exploitation, beyond simply applying known techniques.
1
4
677
stochastic parrot. yeah
The agents also discovered and used at least 10 Linux exploitation techniques that, to the best of our knowledge and all of the experts that we consulted, are novel.
This suggests frontier LLMs can discover and patch new approaches to exploitation, beyond simply applying known techniques.
41
I hope these experts are on @ZackKorman approved experts list that consists of like 3.5 people and a poodle
The agents also discovered and used at least 10 Linux exploitation techniques that, to the best of our knowledge and all of the experts that we consulted, are novel.
This suggests frontier LLMs can discover and patch new approaches to exploitation, beyond simply applying known techniques.
1
63