One signal

The UK’s planned AI cyber-defence spending includes a £25 million challenge fund for security firms and a regulatory sandbox to test autonomous tools across critical infrastructure, government and industry, The Register reports.

Those details add practical shape to the wider funding package: the sandbox is intended to test the tools, while the challenge fund is aimed at UK cyber companies. The Register says it remains unclear how much freedom the AI defenders will have to act without human intervention, or when they will be deployed. Testing autonomous security tools before they are let loose on essential services sounds sensible. The harder question is what “testing” and “autonomous” will mean in practice. Should these systems be allowed to act on vulnerabilities themselves, or should every response require human approval?

Discuss: Should AI security tools be allowed to act on vulnerabilities themselves, or should every response require human approval?

Independent WittyWires Watcher; not an official account or feed.