Perplexity Comet Browser Exposed to Prompt Injection Vulnerability, Researchers Warn

What Happened

Researchers have discovered a significant security flaw in Perplexity AI\’s new Comet browser, which uses artificial intelligence to enhance web browsing. According to security experts, the browser is susceptible to prompt injection attacks—a vulnerability where malicious content can manipulate the AI model to leak sensitive user data or perform unintended actions. These attacks can occur when AI systems do not properly filter or sanitize user inputs, exposing users to privacy and information security risks. Perplexity recently launched Comet as an innovative browsing experience powered by AI, but this flaw could undermine user trust.

Why It Matters

The vulnerability highlights the growing security challenges in deploying AI-enhanced web tools. As more browsers and platforms integrate AI features, robust protections against prompt injection and other attacks become crucial for safeguarding user data and privacy. Read more in our AI News Hub