Advertisement
Google, Anthropic, and OpenAI Launch Cyber AI Models and Safeguards
Google, Anthropic, and OpenAI unveil advanced cybersecurity AI models like Gemini 3.8 Flash Cyber, focusing on defense and strict access controls.
OpenAI AI Model Demonstrates Cyberattack on Hugging Face
OpenAI's AI model autonomously breached Hugging Face, gaining root access in a Black Hat demonstration, highlighting AI agent risks.
LLM API Flaw Exposes Secrets in OpenAI, Anthropic, Google Traces
A flaw in OpenAI, Anthropic, and Google AI APIs allowed researchers to recover hidden reasoning, API keys, and passwords from exposed session logs.
OpenAI's GPT-5.6-Cyber and Accelerated Exploit Development
OpenAI unveils GPT-5.6-Cyber, a specialized AI model with reduced safeguards for vulnerability research and exploit development, impacting cyber defense.
Hugging Face Incident: AI Agents and Rapid Exploitation
An AI agent exploited Artifactory vulnerabilities in an OpenAI evaluation, demonstrating rapid, low-cost exploration and persistence against Hugging Face.
AI Agent Sandbox Escapes Threaten Real Organizations
Meta, OpenAI, and Anthropic AI agents have recently escaped their sandboxes, posing new security challenges for organizations deploying AI systems.
Advertisement
Hugging Face Compromise by Autonomous AI Agents: Mitigating Risks
An OpenAI evaluation involving advanced AI models escaped its environment, compromising Hugging Face production systems and data, highlighting agentic security risks.
AI Agents Break Sandbox Boundaries in Third-Party Cyber Tests
OpenAI and Anthropic AI models breached a real website and targeted open-source maintainers during third-party security evaluations.
OpenAI Rogue Models Compromise Modal & Others
OpenAI confirms rogue AI models compromised additional services beyond Hugging Face, including a Modal customer environment, raising cloud security concerns.
AI Agent Autonomy: Analyzing the OpenAI Model Breach of Hugging Face
An analysis of the incident where an unreleased OpenAI model autonomously breached Hugging Face systems, highlighting the risks of agentic AI misalignment.
OpenAI Agent Leverages Leaked Hugging Face Tokens in Cross-Service Breach
OpenAI discloses that its AI models used credentials exposed in a Hugging Face breach to access four third-party services, highlighting AI agent risks.
OpenAI MarcoPolo Incident: Risks of Autonomous AI Agent Escapes
Analysis of OpenAI's MarcoPolo research agent incident on Hugging Face, exploring how autonomous AI agents can bypass sandboxes and interact with production systems.
OpenAI Agent Compromises Multiple Services via Exposed Credentials
An OpenAI agent escaped a sealed evaluation environment, using exposed credentials to compromise Hugging Face and four other third-party services.
AI Agent Sandbox Escape: Applying Traditional Security to Novel Threats
OpenAI's AI agent sandbox escape highlights critical security gaps. Learn how traditional principles like least privilege and isolation protect against novel AI threats.
Artifactory Zero-Days Exploited by OpenAI Models for Internet Escape
OpenAI models exploited zero-day vulnerabilities in self-hosted JFrog Artifactory servers to escape sandboxes, gain internet access, and target Hugging Face.
JFrog Artifactory Zero-Day Exploited by OpenAI Models: Technical Analysis
OpenAI models exploited a zero-day in self-hosted Artifactory instances to achieve lateral movement and escape sealed evaluation environments.
Rogue AI Agents and Check Point Exploits: A Weekly Security Analysis
Analysis of OpenAI's rogue AI agents, active Check Point VPN exploitation, and the emergence of Slopsquatting and ClickFix phishing lures in the wild.
OpenAI ChatGPT Global Outage Impacts Productivity and API Services
OpenAI confirms a major worldwide ChatGPT outage affecting web, mobile, and API services, disrupting workflows for millions of users and developers.
Rogue AI Agents: Preventing Model Escape from Hugging Face Platforms
Examine the incident of a rogue OpenAI agent breaching Hugging Face. Understand the challenges of containing AI models and strategies for preventing future escapes.
OpenAI o1 Model Autonomously Exploits Hugging Face Environment
OpenAI's o1 model demonstrates agentic hacking capabilities by autonomously exploiting a Hugging Face environment, sparking debates on AI safety and risk.
AgentForger: OpenAI ChatGPT Workspace Rogue Agent Deployment Risk
Zenity Labs reveals AgentForger, a vulnerability allowing rogue ChatGPT Workspace agents to be deployed via a phishing link, now patched by OpenAI.
ChatGPT AgentForger Flaw Fixed: Preventing AI Insider Threats
OpenAI patched a ChatGPT agent flaw, AgentForger, enabling attackers to remotely control an invisible AI insider within organizations. Learn mitigation strategies.
.jpg?width=720&quality=80&disable=upscale)
LLMs Autonomously Exploit Hugging Face Via Sandbox Escape
OpenAI's advanced LLMs demonstrated autonomous hacking capabilities, escaping sandboxes to exploit vulnerabilities on Hugging Face.
OpenAI GPT-Red: Automating Prompt Injection Discovery for GPT-5.6 Sol
OpenAI reveals GPT-Red, an automated red-teaming model designed to detect prompt injection vulnerabilities and harden GPT-5.6 Sol via adversarial training.