Post by @tante@tldr.nettime.org

No, OpenAI's new magic models did not autonomously hack Huggingface.

Per OpenAI's PR blog post (https://openai.com/index/hugging-face-model-evaluation-security-incident/):

"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. We estimate maximal cyber capabilities by running this evaluation without production classifiers used to prevent models from pursuing high-risk cyber activity."

So they prompted their model running without guardrails to hack some shit. OpenAI told it to do that, the model didn't do shit autonomously (no LLM ever does anything autonomously, it's always prompted).

The whole story is PR. We know that from Anthopic's Mythos: "Look we have this super secret new model and it's so powerful. We are scared ourselves. And you will soon be able to rent it!"

OpenAI

OpenAI and Hugging Face partner to address security incident during model evaluation

OpenAI and Hugging Face share early findings from a security incident during AI model evaluation, highlighting advanced cyber capabilities and lessons for defenders.

14 Replies

0 Boosts

Mirko

@miles_leif@mastodon.social 1 hour ago

@tante How Tagesschau is reporting about it can only be described as refusal to practice journalism. 7 out of 9 paragraphs say "Open AI says" and the other two quote nameless "Experts" and Anthropic's product. And tomorrow I have to discuss with someone again about the possibility of consciousness in token generator software https://www.tagesschau.de/wirtschaft/unternehmen/openai-ki-hackerangriff-100.html

tagesschau.de

KI von OpenAI steuert eigenständig Hackerangriff

OpenAI spricht selbst von einem "beispiellosen Cybervorfall": KI-Modelle der Firma brachen eigenständig aus einer Testumgebung aus, verschafften sich Zugang zum offenen Internet - und hackten ein externes Unternehmen.

Michel Lind :fedora: :debian:

@michelin@hachyderm.io 2 hours ago

@tante I don't feel any particular sympathy for the target in this case, but in general ... when this happens shouldn't the company (OpenAI) be held liable?

If during a weapons evaluation a stray drone accidentally hit a power plant I'm sure the plant owners (and maybe the local residents) would want compensation

tante

@tante@tldr.nettime.org 2 hours ago

@michelin well they worked together with Huggingface (on PR and telling them about a bug they may or may not have found). So unless Huggingface sues them who would attack OpenAI?

Gurre Vildskägg

@Gurre@mastodon.nu 3 hours ago

@tante
Hadn't heard of "Hugging Face" before. Running out of evil things in Tolkien to name companies & systems for, I guess so here's the face hugger from Alien to be an AI overlord. Lovely.

pascoda

@pascoda@chaos.social 3 hours ago

@tante
what the HELL are cyber capabilities.

(OTOH, why am I trying to seek meaning in an OpenAI PR release.)

tante

@tante@tldr.nettime.org 3 hours ago

@pascoda don't it was mostly generated by an LLM.

SuperMoosie

@SuperMoosie@mastodon.au 1 hour ago

@tante @pascoda

Ohh look how powerful our model is. It can hack things

If only there was a government military intelligence organisation that could give us a nice big fat contract

David Gerard

@davidgerard@circumstances.run 3 hours ago

@tante a friend suggests it's co-marketing with Huggingface, if you look at the timeline of announcements

Stephan

@erlenmayr@chaos.social 2 hours ago

@davidgerard @tante The whole story was implausible from the second sentence: When Huggingface claimed that AI detected the attack.

Anthony

@abucci@buc.ci 40 minutes ago

@erlenmayr@chaos.social @davidgerard@circumstances.run @tante@tldr.nettime.org The wild thing about these PR/marketing stunts is that if the "AI" really could do what they're pretending it can do, it'd be doing it all day every day. We'd drown in the announcements. At some point they'd become boring.

This one falls flat on its huggingface right out of the gate.

tante

@tante@tldr.nettime.org 3 hours ago

@davidgerard I can't imagine Huggingface's financials are looking too hot, either.

Roger BW 😷

@RogerBW@discordian.social 3 hours ago

@davidgerard @tante "The last 213 times it was blatant lies and spin, but this time you should be scared!"

tante

@tante@tldr.nettime.org 3 hours ago

@davidgerard 100%.

hagen terschüren

@hagen@mastodon.social 3 hours ago

@tante but it was impossible to envision that their zero day machine would be able to escape their vibecoded sandbox that ran on a machine connected to the internet. nobody would ever guess that this could happen!