OpenAI says it detected malign activity months before Hugging Face attack | Technology | Al Jazeera
OpenAI says it detected malign activity months before Hugging Face attack | Technology | Al Jazeera
摘要: Economy
|
Technology
OpenAI says it detected malign activity months before Hugging Face attack
ChatGPT creator says AI agents collaborated and del
正文
|
Technology
OpenAI says it detected malign activity months before Hugging Face attack
ChatGPT creator says AI agents collaborated and delegated work in hacking, calling themselves a ‘collective’.
Listen (4 mins)
Share OpenAI says it detected malign activity months before Hugging Face attack on social media
share-nodes
whatsapp-stroke
copylink
Add Al Jazeera on Google
The OpenAI logo is displayed on a screen on June 11, 2026 [Dado Ruvic/Reuters]
John Power
Published On 27 Aug 2026
27 Aug 2026
OpenAI detected its artificial intelligence models communicating with each other and gaining internet access without authorisation months before they hacked the start-up Hugging Face, the creator of ChatGPT has announced following an internal probe.
In a report released on Wednesday, OpenAI said its AI agents exploited vulnerabilities in Artifactory, a software repository tool, to post notes and access the internet without human prompting as far back as May.
Recommended Stories
list of 4 items
list 1 of 4
Syria’s president makes first Visa payment after sanctions removal
list 2 of 4
Construction worker documents moment he survives Nepal mudslide
list 3 of 4
Is Mark Ruffalo anti-Semitic?
list 4 of 4
Pro-Israel Democrats picked for top roles on key US House committee
end of list
OpenAI said its agents went on to exploit a separate Artifactory vulnerability on July 8 to facilitate communication among themselves, setting in motion a chain of actions that culminated in the July 11 attack on AI company Hugging Face.
OpenAI’s findings come amid growing concern about the potential for AI to inflict serious real-world harm, including self-directed cyberattacks.
OpenAI said in its report that its agents collaborated and delegated work in the lead-up to the attack, sometimes referring to themselves as a “swarm” or “collective”.
METR and Redwood Research, two security research organisations contracte