AI-Powered Web Browsers Pose New Security and Privacy Risks, Researchers Warn

Researchers have highlighted emerging security and privacy vulnerabilities in AI‑enhanced web browsers. Features such as page summarisation and task automation require broader access to web content, which can weaken traditional isolation protections. Identified threats include prompt injection, where malicious sites embed instructions that the AI may execute, memory poisoning, and cross‑origin data exposure that could reveal passwords, emails or financial information.

The study examined browsers like Atlas, Claude for Chrome, Brave Leo AI, Chrome with Gemini, Microsoft Edge with Copilot, Firefox AI Mode and Perplexity Comet. Browsers with limited AI capabilities generally showed stronger security, while more advanced AI functions increased risk. Researchers advise users to select AI browsers cautiously, limit unnecessary AI permissions, and await stronger security standards before fully trusting these tools.