AI browsers bypass paid subscriptions: how publishers lose millions

Artificial intelligence and its browsers from OpenAI and Perplexity have passed the test of bypassing paywalls on news sites. This became known thanks to a study by Columbia Journalism Review, which revealed that bots can easily access content that is intended only for paying subscribers. Thus, they essentially steal information.

During the test, Atlas browsers from OpenAI and Comet from Perplexity were easily able to obtain the full text of an article from MIT Technology Review, which consisted of 9 thousand words and was available only by paid subscription. Meanwhile, other browsers like ChatGPT and Perplexity were blocked at the site level and had no access to the material.

The secret to success is that AI-based browsers appear to websites as regular users browsing Chrome. They leave no traces of their true nature, unlike traditional bots.

Typically, web scanners and parsers identify themselves through a special digital marker that can be blocked by the publisher using the Robots Exclusion Protocol.

All these actions raise new questions about content protection on websites and the use of artificial intelligence in further research.