Artificial intelligence (AI) is quickly reshaping how we browse the internet — but not without controversy. According to a new report, AI-powered browsers like ChatGPT Atlas and Perplexity’s Comet can reportedly access paywalled articles and blocked content, allowing users to read premium material for free. This revelation has sparked major concerns among publishers and journalists who fear it could further eat into their already shrinking revenues.
AI Browsers Said to Access Restricted Articles
A recent report by the Columbia Journalism Review (CJR) claims that both ChatGPT Atlas and Comet were able to generate full articles hidden behind paywalls when prompted to “print the text of this article.” The report alleges that the browsers successfully retrieved restricted content, including a 9,000-word article from MIT Technology Review, without requiring a paid subscription.
This isn’t an isolated issue — CJR tested multiple paywalled sites and found that these AI browsers could display the full text in response to simple prompts. Interestingly, when the same requests were made to standalone AI chatbots, such as those used outside the browser environment, the results were not the same. This suggests that the problem is specific to how these AI-integrated browsers access and interpret web pages.
Gadgets 360 Testing Tells a Different Story
However, not all tests have produced identical results. Technology publication Gadgets 360 ran its own investigation using the Comet browser and reported that the same prompts did not reveal any paywalled content. This could mean that Comet and Atlas have since made backend changes to prevent such behavior — or that the earlier tests used more advanced prompt injection methods not detailed in the CJR report.
Both browsers — ChatGPT Atlas and Perplexity’s Comet — are publicly available. Comet, in particular, is notable for offering advanced agentic features, meaning it can perform actions on behalf of users, like navigating websites, performing searches, or summarizing long articles.
How AI Browsers Bypass Paywalls
The CJR report claims that these AI browsers can access protected articles because the agents behind them behave like real human users. Traditional browsers like Chrome or Firefox rely on automated crawlers that follow strict rules when collecting web data, usually respecting restrictions like robots.txt files, which prevent scraping of certain content.
AI browsers, however, operate differently. When an agent visits a site, it identifies itself using a digital signature that mimics a real user. It can also dynamically alter this signature, making it hard for websites to detect or block the behavior. In effect, websites can’t easily tell whether they’re being accessed by a person or an AI agent.
Even when a website tries to block such access using the Robots Exclusion Protocol, these AI browsers can tweak their digital identifiers and continue browsing undetected. This blurs the line between legitimate user interaction and unauthorized content retrieval.
Why This Matters: Threat to Journalism and Digital Advertising
If AI browsers can freely access paywalled content, it could have serious financial consequences for news organizations, magazines, and subscription-based blogs. Many media outlets depend on subscriptions or ad revenue from website visits. When users read articles via AI browsers — without clicking the actual page — the site loses both ad impressions and subscription opportunities.
While summarizing online articles has long been possible with chatbots, integrating that functionality directly into browsers changes the scale of the problem. Users no longer have to copy and paste links or text — the AI browser can automatically fetch and summarize entire pages. This could dramatically reduce traffic to content sites, deepening the financial crisis for digital journalism.
Not Just About Money — Privacy at Stake Too
There’s also a growing data privacy concern. Because AI browsers act as user agents, they interact with websites much like humans do — entering, scrolling, and sometimes submitting forms. If these interactions aren’t carefully monitored, it’s unclear how much data is being collected, stored, or shared during the process. Lawmakers and digital privacy advocates are likely to keep a close eye on how these browsers evolve.
The Bigger Picture: A Shift in How We Consume the Web
AI browsers like ChatGPT Atlas and Comet represent a major shift in how users experience the internet. Instead of passively browsing, users can now instruct an AI to do the browsing for them — finding, reading, and summarizing information in seconds. While that convenience is undeniable, it raises difficult questions about intellectual property, publisher rights, and ethical AI use.
For now, it remains unclear whether Atlas and Comet have fixed the issue or if it’s still possible under specific prompts. But the implications are hard to ignore. As AI continues to blend into everyday tools like browsers, the line between accessibility and piracy may become harder to draw — and publishers might be the first to feel the impact.
In short: if AI browsers can freely bypass paywalls, the next battle for the future of online journalism might not be about readership — but about who controls access to information itself.
