Security News

US Senators Push for AI Cyber Defences as Attackers Outpace Patching

The Record · 19 Nov 2026
Key Takeaway Plan and test your patching process now so critical fixes can be applied quickly without breaking the systems your business depends on.

Two senior US senators, Maria Cantwell and Mike Rounds, have asked the White House to launch a national effort against AI-enabled cyber threats. In a letter to Director of National Intelligence Jay Clayton, they warned that AI is "rapidly increasing the ability to discover and exploit weaknesses in software" and that national defences "are not keeping pace".

The core problem, they said, is speed. AI can help attackers find and use software weaknesses faster than many organisations can respond. Even when fixes exist, testing and installing patches across interconnected systems can take weeks or months, and may disrupt essential services. The letter points to recent incidents in which AI was used to find obscure bugs and power intrusions in Japan, and to banks that cannot tell whether AI agents accessing accounts are rogue or acting for customers.

The senators want the White House AI task force to speed up defensive AI tools that can continuously identify vulnerabilities, develop and test fixes, and block attacks before they spread. They also want security standards for autonomous AI agents, with CISA leading adoption across banks, hospitals, utilities and government agencies. A "national cyber defense grand challenge" was also proposed to shorten the time from discovering a critical flaw to effective protection.

The letter follows an FBI warning about a Chinese tool that automates large parts of the intrusion process and was used in dozens of attacks in Japan, South Korea and the US. Clayton has not commented on the letter, but said recent outages at major banks, tech companies and airlines are linked to a surge of patching after AI scanners found vulnerabilities faster than ever. Some upgraded systems, he noted, are not returning as expected, including handoffs between systems.

Why it matters

The assumption that you have weeks to test and roll out patches is weakening when AI shortens the time attackers need to find and exploit flaws. The outages Clayton describes show the other side of the trade-off: rushed upgrades can break dependent systems, so patching speed and rollback readiness now need to be planned together. Organisations that let AI agents access accounts or data also face a gap if they cannot distinguish legitimate agents from rogue ones.

What to do

  • Map which of your systems depend on each other so patch handoffs can be tested before rollout.
  • Prioritise critical vulnerability fixes and set a target timeframe shorter than weeks.
  • Prepare rollback plans for upgrades in case systems do not come back as expected.
  • Review which AI agents have access to accounts or data and confirm each acts on someone's behalf.
Running AI agents? Eight controls for securing them, mapped to the ISM ->

Summarised by CISO AI from The Record, written with Claude Sonnet 5.5. We link back to every original so you can read it yourself.