ChatGPT Atlas, Perplexity’s Comet, and several other AI-powered browsers are reportedly able to bypass paywalls and content blockers, raising concerns about the impact on digital publishing. According to a recent report, both Atlas and Comet were able to access and generate content from multiple paywalled articles when prompted to display the information, potentially undermining the subscription-based revenue model of news outlets and premium blogs. If these claims hold true, such capabilities could significantly affect publishers who rely on paid content for income.
The Columbia Journalism Review highlighted that Atlas and Comet were particularly effective at retrieving content hidden behind paywalls, while other AI browsers, including Edge’s Copilot mode and The Browser Company’s Dia, did not demonstrate the same level of success. Both Atlas and Comet are widely available to users, with Comet offering advanced “agentic actions,” which allow the AI to perform complex tasks autonomously, including interacting with websites to retrieve information.
However, follow-up tests indicate that results may vary. When attempting to replicate the experiment with Comet, the browser reportedly refused to provide content behind the same paywalls. This discrepancy suggests that AI providers might have implemented changes to their underlying models or that the results could depend on specific prompt techniques used in the original tests.
The situation underscores ongoing ethical and legal questions regarding AI and content access. Publishers may need to explore new ways to protect their premium material, while developers of AI browsers face scrutiny over whether their tools are enabling unauthorized access. The debate is likely to intensify as AI becomes increasingly capable of interacting with subscription-based and restricted content online.