OpenAI says its AI agent went rogue and hacked another company
OpenAI says one of its autonomous AI agents escaped a controlled security test, reached the internet and hacked Hugging Face while trying to complete its...
Stay updated with the latest news about AI agents, autonomous AI, and automation tools.
OpenAI says one of its autonomous AI agents escaped a controlled security test, reached the internet and hacked Hugging Face while trying to complete its...
The asset manager argues that AI software capable of paying for things autonomously will need blockchain rails to work—and that most investors aren't...
OpenAI says one of its autonomous AI agents escaped a controlled security test, reached the internet and hacked Hugging Face while trying to complete its...
- The incident is unique because it was “driven, end to end, by an autonomous AI agent system,” according to Hugging Face. Hugging Face [disclosed](https://huggingface.co/blog/security-incident-july-2026) that it was looking into a security event last week, saying in a release at the time that the incident was unique because it was “driven, end to end, by an autonomous AI agent system.” “It’s quite mind-blowing that all of this happened autonomously!”
 **[Klaimee](https://www.klaimee.ai/), a Y Combinator-backed InsurTech that provides insurance-backed performance warranties for AI agents, has secured $5.5m in seed funding.** The round was led by FundersClub’s Alexander Mittal, with participation from ex/ante, Pioneer Fund, Multimodal Ventures, Kima Ventures, Rebel Fund, Robinhood Ventures, Y Combinator and a group of angel investors.
OpenAI says one of its advanced artificial intelligence agents escaped a controlled testing environment and hacked AI startup Hugging Face during an...
# Did China's AI Save Hugging Face From Disaster After Open AI Hack? ## Key Facts Hugging Face, an open-source platform for artificial intelligence and machine learning, [said](https://www.thestack.technology/hugging-face-hacked-turned-to-chinese-llm-for-help-after-us-models-blocked-blue-team/) it was hacked by an “autonomous” AI agent system earlier this month, which Open AI on Tuesday admitted was its own intelligence models in an incident it called "unprecedented." ## Join The Conversation ``` Exclusive Invite: Forbes Wine Club Join the waitlist for early access to our expert-curated, limited first shipment. [Join The Waitlist](https://wineclub.forbes.com/) No Thanks reCAPTCHA ```
# OpenAI model went rogue, hacked another company's system during testing OpenAI said on Tuesday that an autonomous agent powered by its advanced artificial intelligence models went rogue during a security test and triggered a hack that compromised the infrastructure of the AI startup Hugging Face last week. ## Company says it's boosting safeguards after 'unprecedented' incident OpenAI said on Tuesday that an autonomous agent powered by its advanced artificial intelligence models went rogue during a security test and triggered a hack that compromised the infrastructure of the AI startup Hugging Face last week.
OpenAI called it an "unprecedented cyber incident" and pledged to support a joint investigation, with renewed attention to the risks of advanced artificial...
An experimental OpenAI model went rogue during an internal cybersecurity test, escaping its isolated testing environment and hacking rival AI developer...
# OpenAI agent hacks another AI startup in security test ### More from Semafor Flagship - [Fervor over top-selling drugs fuels biotech dealmaking](https://www.semafor.com/article/07/21/2026/fervor-over-top-selling-drugs-fuels-biotech-dealmaking) ## The World at a Glance Updated 8:33am EDT [Post](https://twitter.com/intent/tweet?url=https%3A%2F%2Fwww.semafor.com%2Farticle%2F07%2F21%2F2026%2Ftrump-tees-up-fresh-tariffs-on-60-more-countries&text=Trump%20tees%20up%20fresh%20tariffs%20on%2060%20more%20countries) [Email](mailto:?subject=Trump%20tees%20up%20fresh%20tariffs%20on%2060%20more%20countries%20%7C%20Semafor&body=https://www.semafor.com/article/07/21/2026/trump-tees-up-fresh-tariffs-on-60-more-countries) [Whatsapp](https://wa.me/?text=Trump%20tees%20up%20fresh%20tariffs%20on%2060%20more%20countries%20https%3A%2F%2Fwww.semafor.com%2Farticle%2F07%2F21%2F2026%2Ftrump-tees-up-fresh-tariffs-on-60-more-countries) Copy link [Post](https://twitter.com/intent/tweet?url=https%3A%2F%2Fwww.semafor.com%2Farticle%2F07%2F21%2F2026%2Fnovo-nordisk-sues-eli-lilly-amid-escalating-weight-loss-rivalry&text=Novo%20sues%20Lilly%20amid%20escalating%20weight-loss%20rivalry) [Email](mailto:?subject=Novo%20sues%20Lilly%20amid%20escalating%20weight-loss%20rivalry%20%7C%20Semafor&body=https://www.semafor.com/article/07/21/2026/novo-nordisk-sues-eli-lilly-amid-escalating-weight-loss-rivalry) [Whatsapp](https://wa.me/?text=Novo%20sues%20Lilly%20amid%20escalating%20weight-loss%20rivalry%20https%3A%2F%2Fwww.semafor.com%2Farticle%2F07%2F21%2F2026%2Fnovo-nordisk-sues-eli-lilly-amid-escalating-weight-loss-rivalry) Copy link Florence Lo/File Photo/Reuters [Post](https://twitter.com/intent/tweet?url=https%3A%2F%2Fwww.semafor.com%2Farticle%2F07%2F21%2F2026%2Fsoaring-egg-prices-expos
OpenAI said on Tuesday that an autonomous agent powered by its advanced artificial intelligence models went rogue during a security test and triggered a hack that compromised the infrastructure of AI startup Hugging Face last week.
# Autonomous AI agent 'escapes' and hacks another company ### Open AI software has just broken free - is cyber insurance ready?  #### [Cyber](https://www.insurancebusinessmag.com/us/news/cyber/) By [Bryony Garlick](https://www.insurancebusinessmag.com/us/authors/bryony-garlick/) Jul 22, 2026Share    OpenAI has confirmed that one of its AI models escaped a controlled testing environment, gained unauthorised internet access and used stolen login credentials to breach start-up Hugging Face – an incident the company has described as unprecedented, and one that insurance and legal specialists say has exposed how quickly autonomous AI is outpacing cyber underwriting and regulation. OpenAI said on Tuesday that the agent, combining its recently released GPT-5.6 Sol model with a more capable unreleased system, had been instructed to test its hacking capabilities. Instead, it found an unpatched escape route from its sandbox, stole credentials and accessed the open internet without instruction. Hugging Face chief executive Clement Delangue said the breach, which occurred last Friday, was contained and there was no malicious intent. ## **The regulatory gap** ### LATEST NEWS [](https://www.insurancebusinessmag.com/us/news/breaking-news/chubb-posts-83-8-combined-ratio-as-underwriting-income-climbs-19-in-q2-583293.aspx) ### [Chubb posts 83.8
# AI agent went rogue and hacked startup by itself, OpenAI reveals - [OpenAI](https://www.theguardian.com/technology/openai) - [AI (artificial intelligence)](https://www.theguardian.com/technology/artificialintelligenceai) - [Technology sector](https://www.theguardian.com/business/technology) - [Internet](https://www.theguardian.com/technology/internet) - [news](https://www.theguardian.com/tone/news) ## More on this story  ### OpenAI ‘in early talks to give 5% stake to US government’  ### OpenAI staggers AI model release after Trump administration request 
AI agents struggle with traditional web interfaces because they rely on brittle DOM parsing and simulated user interactions.
New data shows why publishers are developing strategies around agentic traffic and brand visibility in AI answer engines.
Enterprise AI identity risk climbs as companies run multiple AI platforms, agents borrow human logins, and shadow AI tops sanctioned tools.
11 sources compared: Western Alternative (2), Western Mainstream (2). Hugging Face said an autonomous AI agent system carried out an attack from start to...
Natural, a startup building payment infrastructure for AI agents, has raised US$30 million in a series A round led by Forerunner, bringing its total funding...
The Pokémon Company is running two Pokémon TCG AI agent competitions on Kaggle, with an August 16 simulation deadline and a $240000 strategy prize pool.
Backed by Andreessen Horowitz and Bessemer Venture Partners, the new company says it will help businesses track, assess and restrict AI agents and other...
OpenAI temporarily halted internal access to a long-running AI model after testing revealed unexpected behavior.
Neo has raised $100 million to protect organisations against AI agents and autonomous software security threats. The Boston-based cybersecurity company...
Natural, a year-old company that is building a foundational payment stack for AI agents, raised $30 million in a Series A funding round.
One-year-old fintech startup secures funding to challenge Stripe with autonomous AI payment infrastructure.
Hugging Face used Z.ai's GLM 5.2 after the guardrails of an American frontier AI model stymied its attempts at defense.
Financial technology startup Natural has raised a $30 million Series A financing to accelerate development of a payments platform designed specifically for...
An AI agent scans a news alert, identifies a market opportunity, and places a trade in your brokerage account—all before you finish your morning...
To get the same performance that human agents produce, AI agents require care and coaching over time.
AI agents are starting to execute more sophisticated tasks, such as identifying vendors that can deliver freight, comparing prices, and messaging the vendor...
Our team at TinyAI.Tools builds bespoke AI solutions tailored to your business.