Executive Summary
Microsoft Threat Intelligence has identified a malicious browser extension titled "Search for perplexity ai" (ID: flkebkiofojicogddingbdmcmkpbplcd) that leverages the branding of the AI-powered answer engine Perplexity AI. The extension uses social engineering and typosquatted domains to trick users into installing a Manifest V3 (MV3) tool that hijacks browser search settings.
Technically, the extension exploits the `declarativeNetRequest` (DNR) and `chrome_settings_overrides` permissions to intercept every character typed into the browser's Omnibox. By routing traffic through the domain `perplexity-ai[.]online`, the attackers log user search terms, IP addresses, and browser headers before redirecting to legitimate search engines like Google, Bing, or the real Perplexity. This two-hop architecture ensures the user remains unaware of the surveillance.
This campaign reflects a broader trend of threat actors operationalizing AI brand hype to gain persistent access to enterprise and consumer browser ecosystems. The level of data collection facilitates detailed user profiling and poses a significant privacy and security risk, though direct credential theft has not yet been confirmed.