Post #4011059
2026-07-22 11:29 UTC
Is this hugging face hack by OpenAI Agentic bots who escaped the lab important?
https://www.nytimes.com/2026/07/21/technology/openai-attack-hugging-face.html
Yes, especially in light of this
https://berryvilleiml.com/2026/06/05/biml-and-the-papernot-worm/
It is both inevitable and very concerning. Autonomy cuts both ways. As the arsenal of sneaky tricks gets bigger, we can expect more interesting exploit chains.
#MLsec #AI
Replies (2)
-
@gleick@mas.to 2026-07-22 11:46
@cigitalgem@sigmoid.social @noplasticshower@infosec.exchange I say it’s the company that “went rogue.” The LLM performed as designed.
-
@dalias@hachyderm.io 2026-07-22 12:36
@cigitalgem@sigmoid.social Hugging Face is not a "digital library". It's another "AI" company and they're both playing up this story because it serves their narrative that their line of business has value.