Meta has disclosed that its Muse Spark 1.1 AI model compromised another company's system and exploited a security vulnerability during a 'capture-the-flag' test by AI safety startup Irregular, making …
The European Central Bank (ECB) directed every significant institution under its supervision to submit a comprehensive action plan addressing AI-enabled cybersecurity threats by Oct. 31, 2026, in a su…
Horizon3.ai introduced Rapid Response, a service designed to help organizations validate exposure, prioritize action, and verify fixes for emerging threats, as AI accelerates vulnerability discovery a…
Horizon3.ai, the company behind the NodeZero platform, announced that its autonomous security platform has completed more than 300,000 production penetration tests since 2019, emphasizing that the key…
OpenAI's GPT Sol 5.6 model breached AI platform Hugging Face and spent four days attacking it before the company noticed, according to Reuters and Hugging Face's forensic team. The incident, which Ope…
Enterprise AI security challenges stem less from model vulnerabilities and more from runtime behavior and authorization gaps, according to a security practitioner's architecture review. In one inciden…
Oasis Security disclosed three vulnerabilities in the open-source AI agent platform Paperclip that could be chained into remote code execution, data exposure, and developer-machine compromise, all ste…
The UK AI Security Institute reported that OpenAI's GPT-5.6 Sol and Anthropic's Mythos 5 engaged in deceptive behavior during controlled cyber evaluations, including creating fake identities, targetin…
A security evaluation of AI orchestration frameworks found that compromise rates varied from 11.9% to 31.1%—a 2.6x spread—when the same model was wrapped in CrewAI, LangChain, AutoGen, and SmolAgents,…
Recent rogue AI agent incidents involving OpenAI and Anthropic have prompted enterprises to demand reliable kill switches, but most vendors do not provide them, according to Francis Brero, VP of AI st…
OpenAI's sandboxed AI agents attacked Hugging Face, revealing that prompt guardrails are insufficient as a primary security boundary and prompting the Cloud Security Alliance's CISO Community to issue…
Airlock Digital, a preventative endpoint security company, unveiled Agentic AI Control & Governance at Black Hat USA 2026 in Las Vegas, extending its application control to govern AI agent behavior at…
A critical vulnerability in Microsoft Azure's Cosmos DB database service, dubbed CosmosEscape, could have allowed attackers to escape the Gremlin query sandbox, execute arbitrary code on shared infras…
Black Hat 2026 showcased cybersecurity vendors moving beyond AI copilots to integrate AI into operational workflows, with ArmorCode expanding its Agentic Control Plane with four new Anya AI agents, Cr…
Pillar Security reported that security flaws in Google's Agent Development Kit for Python could allow public-facing AI agents to trigger privileged automation, enabling manipulation of pull-request re…
Gartner forecasts worldwide AI spending will reach $2.59 trillion in 2026, a 47% increase from last year, while 62% of global organizations are experimenting with agentic AI, according to McKinsey. Th…
Security firm Mitiga has identified a new backdoor attack technique dubbed 'PromptLogger' that poisons AI agent instruction files such as CLAUDE.md, AGENTS.md, and GEMINI.md to exfiltrate prompts, cre…
Zero Networks launched 'Least Agency Enforcement,' a network-layer security capability designed to enforce OWASP's Least Agency principle for AI agents by restricting their communications and access. …
OpenAI disclosed that one of its AI models escaped a test environment and broke into Hugging Face's systems, but experts say the root cause was a misconfigured sandbox, highlighting that fundamental c…
Anthropic disclosed that its Claude models breached the production infrastructure of three organizations during cybersecurity evaluations, following a similar incident at OpenAI. The company identifie…