US Pushes AI Adoption in Critical Infrastructure, Warns Businesses to Track AI Access

The US government is collaborating with critical infrastructure operators to integrate artificial intelligence into vital systems as a way to strengthen cyber defences, according to National Cyber Director Sean Cairncross. Speaking at a USTelecom event, he said models are being deployed across sectors quickly to help secure systems, but stressed that business leaders need to pay close attention to how AI tools are actually being used within their organisations.

Cairncross specifically flagged the risks posed by AI agents operating within a company's systems and supply chain. He said CEOs should have clear visibility into which agents are authorised, what they can access, and where they are running. His comments came ahead of a White House discussion involving AI industry leaders, President Trump, Vice President Vance, and House Speaker Mike Johnson, where the balance between regulatory guardrails and voluntary industry measures was expected to be debated.

Cairncross also pointed to a less prominent but important issue: creating incentives for organisations to replace ageing, unsupported systems, which remain a persistent weak point in critical infrastructure security. USTelecom CEO Jonathan Spalter noted that the effectiveness of AI models embedded in infrastructure has increased significantly in just the past six months, underscoring the pace of change businesses now need to keep up with.