Close Menu
Daily Guardian
  • Home
  • News
  • Politics
  • Business
  • Entertainment
  • Lifestyle
  • Health
  • Sports
  • Technology
  • Climate
  • Auto
  • Travel
  • Web Stories
What's On

The $200 Popcorn Popper That Keeps Selling Out

October 1, 2026

Challenged Athletes Foundation Receives Rings of Gold Award from U.S. Olympic & Paralympic Committee

October 1, 2026

CleanChoice Energy Closes Approximately $166M in Project Financing and Tax Equity for Two Solar Projects

October 1, 2026

Fabletics Launches Fabletics Reserve Credit Card, Expanding Value and Loyalty Perks to Its More than 2.7 Million Members

October 1, 2026

First Merchants Bank CEO Mark Hardwick to retire at the end of 2026 and President Mike Stewart to become President and CEO

October 1, 2026
Facebook X (Twitter) Instagram
Finance Pro
Facebook X (Twitter) Instagram
Daily Guardian
Subscribe
  • Home
  • News
  • Politics
  • Business
  • Entertainment
  • Lifestyle
  • Health
  • Sports
  • Technology
  • Climate
  • Auto
  • Travel
  • Web Stories
Daily Guardian
Home » OpenAI says it accidentally hacked Hugging Face with a new AI system
Technology

OpenAI says it accidentally hacked Hugging Face with a new AI system

By News RoomJuly 21, 20262 Mins Read
OpenAI says it accidentally hacked Hugging Face with a new AI system
Share
Facebook Twitter LinkedIn Pinterest Email

OpenAI says its AI models mistakenly breached open-source AI platform Hugging Face during internal testing. In a blog post on Tuesday, OpenAI writes that GPT-5.6 Sol and “an even more capable pre-release model” discovered vulnerabilities within their sandboxed testing environment, allowing them to gain access to the internet and target Hugging Face.

On July 16th, Hugging Face disclosed a security incident that it says was driven by “an autonomous AI agent system.” Hugging Face’s AI agents detected and stopped the breach, which OpenAI has now admitted occurred during an evaluation of its models’ cybersecurity capabilities. OpenAI says “all evidence suggests that the models were hyperfocused on finding a solution for ExploitGym,” a benchmark system that measures whether AI models can turn security vulnerabilities into exploits.

As part of efforts to complete the evaluation, the AI models gained access to the internet by exploiting a zero-day vulnerability in the sandboxed environment. From there, OpenAI says its models “inferred that Hugging Face potentially hosted models, datasets and solutions for ExploitGym,” and then “searched for and successfully found ways to gain access to secret information that it could use to cheat the evaluation:”

In one example, the model chained together multiple attack vectors, including using stolen credentials and zero-day vulnerabilities to find a remote code execution path on the Hugging Face servers.

But as serious as this incident is, OpenAI appears to be using the “unprecedented” attack as an opportunity to make its AI systems look good — especially as it competes with cybersecurity rivals, like Anthropic’s Mythos and Gemini Flash 3.5 Cyber. OpenAI’s blog post has a chart showing how GPT-5.6 Sol is getting better at sustaining multi-step cyber operations, and also encourages enterprise customers to sign up to access its “Cyber” security model.

OpenAI adds that it’s now working with Hugging Face to investigate the security incident, and will implement new controls within its research environment.

Share. Facebook Twitter Pinterest LinkedIn Tumblr Email

Keep Reading

Huawei made a whole camera to attach to its latest phone

Vivo’s X Fold 6 accidentally feels like a throwback

Elon Musk’s Grokipedia has a ‘newly refreshed’ design

All the latest news on Meta’s cute, creepy Muse AI agent

‘Xbox is not for sale’ says Microsoft’s gaming chief

Here’s what AI leaders are saying about Trump’s new safety plan

Google announces Gemini 4 and says it’s so capable that only ‘trusted cyber defenders’ can have it right now

Amazon’s delivery driver smart glasses will reportedly take photos ‘almost constantly’

Reddit says it has to cut back access to ‘Old Reddit’ because of AI bots

Editors Picks

Challenged Athletes Foundation Receives Rings of Gold Award from U.S. Olympic & Paralympic Committee

October 1, 2026

CleanChoice Energy Closes Approximately $166M in Project Financing and Tax Equity for Two Solar Projects

October 1, 2026

Fabletics Launches Fabletics Reserve Credit Card, Expanding Value and Loyalty Perks to Its More than 2.7 Million Members

October 1, 2026

First Merchants Bank CEO Mark Hardwick to retire at the end of 2026 and President Mike Stewart to become President and CEO

October 1, 2026

Latest News

Primech Holdings (Nasdaq: PMEC) Unveils Primech Vision, Integrating a Multi-Robot Facility Fleet and Workforce on One Platform; Live Customer Demonstrations Begin October 2026

October 1, 2026

Police seek answers in crash that killed 2 elders, hurt 2 others on BC First Nation

October 1, 2026

An AI agent tried to hack Canadian government website, researchers say

October 1, 2026
Facebook X (Twitter) Pinterest TikTok Instagram
© 2026 Daily Guardian Canada. All Rights Reserved.
  • Privacy Policy
  • Terms
  • Advertise
  • Contact

Type above and press Enter to search. Press Esc to cancel.

Go to mobile version