Atos UK and Palo Alto Networks Launch a New Service to Defend Against AI Powered Cyberattacks

Atos UK&I has expanded its partnership with Palo Alto Networks to launch a new service built for a threat that barely existed a couple of years ago: cyberattacks powered by frontier AI models. The service, reported on 28 October 2026, is aimed at organisations that need to understand how exposed they are to AI driven attacks and then do something practical about it.

The timing is not hard to explain. According to TechMarketView, frontier AI models have moved fast. Less than a year ago, the report notes, these models could barely complete beginner level cyber exercises. Now the latest models can finish simulated end to end network takeovers, and some can find and exploit vulnerabilities in live infrastructure. That shift changes the maths for defenders. An attacker that once needed a skilled team and plenty of time can now get much further, much faster, with an AI system doing the heavy lifting.

What the new service actually does

The offering has three parts, and each one answers a different question a security team would ask.

The first part is an exposure assessment. Using Palo Alto Networks exposure assessment capabilities, it maps how and where an organisation’s critical systems could be targeted, and sets out practical steps to reduce that risk. This is the “where are we vulnerable” stage, and it matters because most organisations cannot fix everything at once. Knowing which systems an AI powered attacker would reach first helps teams spend their effort where it counts.

The second part is vulnerability management with AI doing the prioritising. Atos’ service uses AI to rank fixes, checking each finding against asset criticality, real world exploitability, ownership and organisational context. In plain terms, a flaw in a system that holds customer data and faces the internet should jump the queue ahead of a flaw in an isolated internal tool, even if both look equally severe on paper. The idea is to stop teams drowning in long lists of vulnerabilities and instead fix the ones an attacker would actually use.

The third part is the round the clock defence: a security operations centre combined with security orchestration, automation and response, plus managed extended detection and response. It is built on Palo Alto’s Cortex platform, hosted in Atos’ sovereign UK data centres and delivered only from the UK by security cleared staff. For organisations in regulated sectors, that UK only delivery is a selling point in itself, since it keeps sensitive security operations inside national borders and in cleared hands.

Why humans still sit at the centre

One detail stands out in how the service is described. While AI handles detection and triage, human analysts keep responsibility for governance, escalation and strategic risk assessment. That division of labour is deliberate. AI is good at sifting through huge volumes of alerts at speed, but deciding how to respond to a serious incident, especially in a hospital, a bank or a government department, still needs judgement about context, consequences and risk appetite.

It also answers a worry that has grown alongside AI security tools: if defenders hand too much to automated systems, a wrong automated response could disrupt essential services. Keeping people in charge of the big calls is meant to guard against that.

Part of a wider UK push

The launch follows the Sovereign MXDR service Atos launched in March, and it lands in the middle of a wider UK push on AI security. Just days earlier, the UK government set out a large AI security package that included funding for a National Cyber Shield and extra money for the AI Security Institute. The Atos and Palo Alto service is a commercial product rather than a government programme, but it points in the same direction: the UK treating AI powered cyber threats as a mainstream concern rather than a niche research topic.

For British businesses, the practical message is fairly simple. AI is now on both sides of the cyber fight. Attackers are already using it to find weaknesses faster, and defenders need tools that can keep up. Services like this one are an attempt to close that gap, pairing AI speed with human oversight and UK based delivery.

Whether it works in practice will depend on results rather than announcements. Organisations that take it up will want to see faster detection, fewer critical vulnerabilities left open and incidents handled without disruption. But the launch itself is a useful marker of where the market is heading: AI security is no longer a futuristic add on. It is becoming part of the standard toolkit.

Frequently Asked Questions

What is frontier AI security?

Frontier AI security is about defending against attacks that use the most advanced AI models. These models can automate steps that once needed skilled human hackers, such as finding vulnerabilities and chaining them into a full network intrusion, so defences need to spot and stop AI driven activity specifically.

What does the Atos and Palo Alto service include?

It includes an exposure assessment to map where critical systems could be targeted, AI powered vulnerability management to prioritise fixes, and a managed detection and response service built on Palo Alto’s Cortex platform, hosted in Atos’ UK data centres and delivered from the UK by security cleared staff.

Why does UK based delivery matter?

For organisations in regulated sectors, having security operations delivered only from the UK by security cleared staff helps meet data sovereignty and compliance requirements, and keeps sensitive incident information inside national borders.