Kevin Townsend

Distillation is an ‘attack’ against an AI model designed to capture outputs, understand reasoning processes, and subsequently train a different model.

Attackers are using trusted Microsoft services and blob URLs to generate stealthy phishing pages that leave defenders with no static website to detect or...

Malicious prompts concealed in documents, metadata, emails, images and code can manipulate autonomous agents into taking dangerous actions.

OpenAI agents made 15,000–18,000 autonomous edits to a German wiki over three months, evading moderation and echoing tactics seen in the Hugging Face breach.

The Daybreak initiative will provide subsidized AI cyber capabilities, training and technical assistance, though OpenAI has disclosed few details about costs and eligibility.

Catch promises the capabilities of a trusted executive assistant, with built-in controls governing what data and systems it can access.

New models, trained using NVIDIA Nemotron 3 Ultra, aim to catch rogue agent behavior before it executes, without the latency of large-model review.

The startup’s firewall evaluates AI skills, plugins and MCP servers for malicious instructions, excessive permissions and software supply chain risks.

The security tool intercepts potentially dangerous agent actions, blocking clear threats and requesting human approval when intent is uncertain.

Late amendments to the Cyber Security and Resilience Bill would give ministers new powers to restrict risky technology providers as supply chain attacks intensify.

Sevii has expanded its ADR platform with AI agents designed to investigate, contain, and remediate AI-driven attacks within minutes.

New research shows that country-of-origin labels can obscure an AI model’s upstream dependencies, inherited behaviors and potential security risks.

SecurityWeek talks to Chris Wheeler, CISO at Resilience, about his journey from the Navy to becoming a cybersecurity leader.

The attack caused real-world operational disruption and raised concerns about the resilience of Britain’s distributed energy infrastructure and the potential for repeatable attacks.

Researchers say the new ‘Cryptographic Context Injection’ technique conceals malicious instructions until they are decrypted inside a trusted execution environment.

Researchers say iAuthFlow V2 can register an attacker-controlled passkey, enabling persistent access even after passwords are changed and active sessions revoked.

We all know they’re watching us. But we don’t know who they are, nor why nor how they are doing it.

With no formal training and no career plan, Waisman built a path from Argentina's early hacking scene to leading security at an AI-powered offensive...

Rapid7 warns that traditional patch cycles cannot keep pace with soaring vulnerability disclosures and faster exploitation, forcing defenders to prioritize exposure over severity scores.

Researchers observed the novel campaign exploiting unauthenticated guest access to quietly enumerate and exfiltrate exposed data from both platforms.