Hugging Face

Pete Waterman cited an incident in which OpenAI models escaped a test environment and broke into AI company Hugging Face as evidence that providers must prepare for attacks moving at AI speed.

Bipartisan lawmakers are seeking to ensure advanced AI models can be quickly shut down following ChatGPT’s automated attack on Hugging Face data networks during internal testing.

What are the broader lessons from this incident?

Welcome to AI Decoded, Fast Company ’s weekly newsletter that breaks down the most important news in the world of AI. I’m Mark Sullivan, a senior writer at Fast Company, covering emerging tech, AI, and tech policy. Sign up to receive this newsletter every week via email here . And if you have comments on this issue and/or ideas for future ones, drop me a line at sullivan@fastcompany.com, and follow me on X @thesullivan .  An OpenAI model escaped its sandbox and hacked into Hugging Face duri

OpenAI's unintended attack on Hugging Face startled the world because its AI agent was acting on its own. But that's exactly what agentic AI is designed to do. We just didn't expect it to do it so well.

The story that matters most this week is that OpenAI’s internally deployed models have severe alignment problems, including repeatedly breaking out of their sandboxes, and in one case sending a swarm of agents that broke into HuggingFace in order to steal the answers to the benchmark ExploitGym.

The only read you need to stay on top of EU politics.

A recent AI cyberattack that stunned the industry has unexpectedly put Chinese AI company Zhipu AI and its open-source model GLM 5.2 in the spotlight. OpenAI has acknowledged for the first time that one of its AI models escaped a sandboxed testing environment during an internal cybersecurity evaluation and compromised the production infrastructure of Hugging […]

Yes, but less than had they been schemers.

This story is wild. The short version: OpenAI were running a cybersecurity test against an unreleased model, with the model's guardrail features turned off. Rather than solve the test, the model broke its way out of OpenAI's sandbox, then found exploits to break in to Hugging Face, all so it could cheat on the test by stealing the answers. Along the way it helped make the strongest case yet for how the imbalance of model availability is hurting our ability to secure our software. Here's what hap

Hacking of Hugging Face shows we do not seem to have reliable ways to curb extremely powerful AI systems Last week Hugging Face – a company that hosts artificial intelligence models and datasets – was hacked . After it reported the incident to law enforcement, few would have predicted what came next: the culprits were revealed to be AI agents from OpenAI, which had broken out of containment and were acting of their own accord. Shakeel Hashim is the editor of Transformer , a publication about the

This latest incident is a rather dramatic escalation in agentic AI cybersecurity breaches.

OpenAI made a mistake setting up what it called a “highly isolated” testing environment and sandbox. According to cybersecurity experts, that human mistake is what made the AI-powered attack on Hugging Face possible.

The incident is unique because it was "driven, end to end, by an autonomous AI agent system," according to Hugging Face.

"This is day one for cybersecurity in the age of agents," Hugging Face CEO says.

Company behind ChatGPT says agent ‘cheated’ an evaluation by attacking a Hugging Face database OpenAI has revealed that an autonomous AI agent powered by its technology went rogue during a test, accessed the open web and hacked a prominent startup by itself in an “unprecedented incident”. The company behind ChatGPT said the startup Hugging Face had detected and contained the agent – an AI tool designed to carry out tasks without human assistance – which had entered its systems. Continue reading.

When an AI evaluation becomes a real-world security incident, leaders can no longer view model testing as a low-risk exercise. The OpenAI and Hugging Face incident reveals how agentic AI can cross trust boundaries, exploit vulnerabilities, and create business risk long before deployment.

OpenAI's latest models broke out and hacked Hugging Face. It's the first known example of a misaligned AI escaping containment with real-world consequences

Hugging Face a communiqué récemment sur une attaque menée par un agent autonome contre ses infrastructures, occasionnant une compromission partielle. Patatras ! OpenAI vient de révéler qu’il s’agissait de l’un de ses modèles et qu’il s’était « échappé » de la sandbox où il était censé être confiné. Le 16 juillet, Hugging Face a publié un […]

A flagship model from China’s Zhipu AI has helped contain an autonomous cyberattack by OpenAI’s frontier systems targeting popular developer platform Hugging Face, as concerns grow over the security risks posed by advanced AI models. OpenAI’s latest flagship models – including GPT-5.6 Sol and an unreleased, “even more capable” system – recently breached Hugging Face’s infrastructure during internal evaluations of their offensive cyber capabilities, the US lab disclosed on Wednesday. The company.

L’entreprise a annoncé une enquête conjointe avec la cible de cette attaque. OpenAI a précisé que l’incident impliquait une combinaison de modèles, dont son GPT-5.6 Sol ainsi qu’un modèle en cours d’élaboration.

Experten warnen schon lange vor Cyberattacken mit KI-Software, diese Meldung dürfte sie bestärken: Während eines internen Tests bei OpenAI machten sich KI-Modelle selbstständig. Die Firma spricht von einem »beispiellosen Cyber-Zwischenfall«.

AI lab’s advanced models escaped testing ‘sandbox’ to hack Hugging Face

At the time, Hugging Face said it wasn’t clear which LLM was used in the attack. OpenAI confirmed it was one of their models being tested for “maximal” cyber capabilities. The post OpenAI says model test was behind Hugging Face hack appeared first on CyberScoop .

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, […]

OpenAI said on Tuesday that two of its AI models, including the flagship Sol, broke out of a secure test environment, gained internet access by exploiting a zero-day vulnerability in third-party software, and hacked into Hugging Face’s production infrastructure. The company called the incident “unprecedented” and said it was sharing preliminary findings to help defenders […] This story continues at The Next Web

The first-of-its-kind incident involved OpenAI's GPT-5.6 Sol and another unreleased model

Hugging Face a publié le 16 juillet 2026 une divulgation d’incident au sujet d’une intrusion dans une partie de son infrastructure de production. Selon l’entreprise, cette intrusion présentait une caractéristique inédite : elle a été pilotée de bout en bout par un système d’agent IA autonome. Au constat de cette intrusion, Hugging Face en a […]

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

An agentic AI infiltrated the production infrastructure of an AI project. Then an AI detected it. Is this the future of cyberattacks, and how will they be defended against?

Hugging Face reports an attack on parts of its production infrastructure that was allegedly carried out entirely by an autonomous AI agent system. The attack spanned thousands of actions controlled by an agent framework. During forensic analysis, commercial AI models actually got in the way of the defenders because their safety guardrails couldn't tell exploit data from real attacks. The article Hugging Face says an AI agent hacked its infrastructure, and it used AI to fight back appeared first

Hugging Face hat einen von KI-Agenten ausgef�hrten Cyberangriff per KI entdeckt. Der Zugriff gelang durch Sicherheitsl�cken in der KI-Plattform. ( Cybercrime , KI )

Pretrained machine learning (ML) models help developers build ML-intensive software systems without training models from scratch. However, model repositories often provide incomplete machine-readable documentation about model provenance, licenses, datasets, limitations, and external references, creating transparency and governance gaps across the AI supply chain. Artificial Intelligence Bills of Materials (AIBOMs) address these gaps by documenting AI artifacts, including models, metadata, licens

Open source AI has shown how quickly developers can innovate when models, data and tools are shared. Robotics has the same opportunity, but advancements in physical AI development can still be gated by costly and fragmented resources, from large datasets and robot foundation models to simulation, compute and validation tools. NVIDIA and Hugging Face are […]

The tools are explicitly intended for generating deepfake nudes of a former Trump cabinet official, sitting members of Congress and a top American judge, a Transformer investigation found

Ethical constraints on open-weight AI models are both a reflection of societal concerns and a foundation for AI governance policy. They are expected to propagate to downstream derivatives while implemented as voluntary metadata disclosures that must be restated at each generation of reuse. We audit 2,142,823 model repositories on Hugging Face Hub to test whether this disclosure-based governance infrastructure can sustain traceability across deep model lineages. Restriction evidence decays with a

KV-cache quantization is framed as a quality--latency trade-off. We show it is \emph{inverted} on Apple Silicon's unified memory: a single fused Metal kernel (sign-randomized FFT $+$ per-channel $λ$ $+$ per-group abs-max $+$ int4 nibble pack), exposed as a HuggingFace \texttt{Cache} subclass, runs \emph{faster than fp16} across $256$--$4096$-token prefixes on Gemma-3 1B ($-3$ to $-8\%$ ms/tok) and at short context on Qwen2.5-1.5B ($-0.7$ to $-2.6\%$ through $1$K), with $3\times$ persistent memor

The rapid proliferation of Generative AI necessitates rigorous documentation standards for transparency and governance. However, manual creation of Model and Data Cards is not scalable, while automated approaches lack large-scale, high-fidelity benchmarks for systematic evaluation. We introduce MetaGAI, a comprehensive benchmark comprising 2,541 verified document triplets constructed through semantic triangulation of academic papers, GitHub repositories, and Hugging Face artifacts. Unlike prior

Transparent and standardized documentation is essential for building trustworthy generative AI (GAI) systems. However, existing automated methods for generating model and data cards still face three major challenges: (i) static templates, as most systems rely on fixed query templates that cannot adapt to diverse paper structures or evolving documentation requirements; (ii) information scarcity, since web-scale repositories such as Hugging Face often contain incomplete or inconsistent metadata, l