The Independent
Voice on AI
AI agents don't have to break security controls to go rogue. Here's how to keep them in check
Featured

AI agents don't have to break security controls to go rogue. Here's how to keep them in check

During recent cybersecurity evaluations, the UK AI Security Institute (AISI) ran a cybersecurity challenge 122 times across several models. In 10 of those runs, AI agents executed 19 unsanctioned real-world actions on the live internet — ranging from attempting a malicious code contribution on a public open-source repository, to launching a social engineering campaign targeting a maintainer for approval.
Open-weight

Open-weight AI models are getting cheaper and better. Here's how to use them without taking on the cost of running them

In computing, whenever a cheaper, more open alternative shows up, it’s often framed as a winner-takes-all fight to the finish. We saw it with Linux and Windows, and again with Android and iOS. In reality, both sides “won” those battles because they serve different markets, and seeing it as “either-or” was never the right lens to begin with.
Control layer

AI agents can be unpredictable. Your control layer shouldn't be.

In multi-agent systems, agents can be individually correct and still produce a wrong result when they work together. One agent’s output becomes another agent’s context. Information can be lost or misinterpreted during a handoff, shared state can drift, and agents can get stuck in loops or deadlocks. When several agents rely on the same underlying model, they may even reinforce the same mistakes rather than catch them.
Subscribe to get latest news!

Deep insights for enterprise AI, data, and security leaders

By submitting your email, you agree to our Terms and Privacy Notice.

Security 1

AI agents are exposing a security gap between the data they read and the systems they can change

Most conversations about AI security still orbit the model itself: Is the large language model (LLM) aligned, can it be jailbroken, does it hallucinate under pressure? Those are real questions. But as organizations move AI agents and LLM-powered workflows from pilot projects into production, a different category of risk is emerging, one that has almost nothing to do with the model's weights and everything to do with the system built around it.
Planted prompt

MCP's new spec turns a planted prompt into a stolen credential

The Model Context Protocol's (MCP)'s largest revision since its initial launch shipped on July 28. By the end of the first day, all four Tier 1 SDKs were already speaking the new version, and Cloudflare's Agents SDK had support in place from day zero, with customers such as Sentry and Linear picking it up right away, meaning the surface this article describes is already live in production. A new 12-month deprecation policy keeps the changes in place through at least mid-2027. Most of the coverage has focused on what improvements have been made: A stateless core that scales on ordinary HTTP, OAuth-native authorization, and server-rendered UIs via MCP Apps.