Cisco tested eight major open-weight artificial intelligence models and found multi-turn jailbreak attacks succeeded nearly ...

Is Dutch Sec. Gijs Tuinman alluding to a European effort to continue using their F-35 jets even if the U.S. stops supporting ...

The moment an AI system can read internal systems, trigger workflows, move money, send emails, update records or approve actions, the risk profile changes.

Attacks against modern generative artificial intelligence (AI) large language models (LLMs) pose a real threat. Yet discussions around these attacks and their potential defenses are dangerously myopic ...

RoguePilot flaw let GitHub Copilot leak GITHUB_TOKEN, while new studies expose LLM side channels, ShadowLogic backdoors, and promptware risks.

Some cybersecurity researchers say it’s too early to worry about AI-orchestrated cyberattacks. Others say it could already be ...

AI agents may work smarter than chatbots, but with tool access and memory, they can also leak data, loop endlessly or act ...

Background In early 2026, OpenClaw (formerly known as Clawdbot and Moltbot), an open-source autonomous AI agent project, quickly attracted global attention. As an automated intelligent application ...

In machine learning, privacy risks often emerge from inference-based attacks. Model inversion techniques can reconstruct ...

In a report published on February 12, ahead of the Munich Security Conference, Google Threat Intelligence Group (GTIG) and Google DeepMind shared new findings on how cybercriminals and nation-state ...

Vitalik Buterin and Davide Crapis, the head of AI at the Ethereum Foundation, are proposing a new system to improve privacy when using large language models.

AI Safety Connect at India AI Impact Summit spotlights guardrails, jailbreaking risks, and the push for stronger AI security ...