UK utilities warned of growing AI-assisted cyber threats

Hackers working for hostile states including Russia and Iran could use AI to help disrupt British water and energy supplies, according to Phil Tonkin, field chief technology officer at Dragos and a former National Grid security chief. AI tools are lowering the specialist knowledge needed to understand operational networks, identify exposed equipment and locate potential points of disruption.

Tonkin said AI is not currently independently carrying out attacks against critical infrastructure. However, Dragos disclosed at a US Senate hearing that an AI agent inside a water utility independently sought computers controlling pumps and attempted around 2,800 passwords. All attempts failed. Remotely connected equipment, particularly at smaller utility sites, can create vulnerabilities when connections are poorly secured.

Tonkin urged operators to identify weaknesses, segment operational networks and improve monitoring rather than rely on a single protective perimeter. He said established security controls need consistent application across both IT and operational systems.

The Cyber Security and Resilience Bill, currently before the Lords, would strengthen incident reporting and supply-chain protections while allowing ministers to intervene against hostile-state involvement in critical infrastructure supply chains. Tonkin welcomed its broader scope but said effectiveness would depend on funding and implementation. He also stressed that growing awareness should not create unnecessary alarm, noting longstanding electricity-sector investment in cybersecurity and resilience.