Perplexity, a rapidly emerging AI search tool, is currently under scrutiny for serious allegations that extend beyond concerns about accuracy and hallucinations. These new challenges involve pressing issues regarding user privacy and data handling, which could lead users to reconsider their assumptions about the confidentiality of their interactions with AI.
A recently filed class-action lawsuit from an anonymous user, referred to as John Doe, has raised troubling questions about the company’s data handling practices. The suit alleges that Perplexity’s so-called “incognito mode” is misleading and does not adequately protect user privacy.
Is Perplexity’s “incognito mode” actually private?
The lawsuit claims that sensitive conversations—ranging from financial advice to health-related inquiries—were shared with third parties like Google and Meta. Reports indicate that this data was shared even when the user had selected the incognito mode, which is intended to minimize tracking and data collection.
Worryingly, the lawsuit alleges that various types of data—including IP addresses, email addresses, geolocation, and even complete chat transcripts—may have been transmitted for advertising purposes. Furthermore, it accuses Perplexity of using tracking tools similar to those found in online marketing without providing clear notifications to users. In some instances, the lawsuit claims entire chat histories could be publicly accessible via easily obtainable links.
Why this lawsuit could change how we trust AI
This situation transcends a single application; it touches on a broader issue regarding trust in AI technologies. AI tools often feel personal, making users prone to oversharing. The lawsuit claims extensive user data has been shared with advertising giants, and it highlights that Perplexity does not make its privacy policy as transparent as its competitors.
If the allegations prove accurate, this case could initiate a demand for greater transparency across AI platforms, prompting users to be more cautious about their interactions. While these are currently just claims, they serve as a reminder that the next time you interact with an AI tool, the information you share might not be as secure as you think.
