This latest security issue highlights the challenges of security in an AI era.

A researcher has found a major vulnerability in NLWeb, Microsoft's underlying protocol for letting AI agents talk to websites ...

Microsoft's new AI-powered NLWeb protocol, designed to enhance websites with ChatGPT-like capabilities, has already ...

AI has prevented real-world cyberattacks before they could begin, proving AI fast enough to beat human threat actors to ...

It’s a protocol that’s supposed to be ‘HTML for the Agentic Web,’ offering ChatGPT-like search to any website or app…The flaw allows any remote users to read sensitive files, including system ...

OpenAI fixed the exploit, but it used a malicious prompt hidden in a Google Drive file to trawl through their personal ...

The bug, tracked as CVE-2025-31199, could allow hackers to access files in the Downloads folder, as well as caches utilized ...

Microsoft’s Threat Intelligence team has identified a now-fixed security vulnerability in Apple’s macOS Spotlight search ...

“It was the first reverse engineer at Microsoft, human or machine, to author a conviction case — a detection strong enough to ...

Google's AI-powered tool designed to find bugs, Big Sleep, has reported its first batch of 20 security vulnerabilities in ...