The emergence of AI-powered browsers represents a significant shift in how artificial intelligence interacts with web content.
However, it has also introduced unprecedented challenges for digital publishers and content creators. Last week, OpenAI released Atlas, joining a growing wave of AI browsers including Perplexity’s Comet and Microsoft’s Copilot mode in Edge, that aim to transform how people interact with the web.
Unlike traditional browsers such as Chrome or Safari, these AI browsers possess “agentic capabilities”—sophisticated tools designed to execute complex, multistep tasks autonomously.
However, their ability to seamlessly bypass paywalls and content restrictions by mimicking legitimate human users has raised serious concerns about intellectual property protection and content monetization across the digital publishing industry.
AI browsers present fundamentally new problems for media outlets and publishers because agentic systems make it increasingly difficult for content creators to know and control how their articles are being accessed and utilized.
When researchers tested Atlas and Comet against a nine-thousand-word subscriber-exclusive article from the MIT Technology Review, both browsers successfully retrieved the full text.
Notably, when the same request was issued in ChatGPT’s and Perplexity’s standard interfaces, both systems responded that they could not access the content because the Review had blocked the companies’ crawlers.
The critical distinction lies in how these AI browsers operate. To a website, Atlas’s AI agent appears indistinguishable from a person using a standard Chrome browser.
When automated systems like crawlers and scrapers visit a website, they identify themselves using a digital ID that tells the site what kind of software is making the request and its purpose.
Publishers can selectively block specific crawlers using the Robots Exclusion Protocol—a standard defense mechanism that many outlets have implemented.
However, because AI browsers like Comet and Atlas appear in site logs as normal Chrome sessions, blocking them risks also preventing legitimate human users from accessing a site.
This fundamental technical limitation makes it exceptionally difficult for publishers to detect, block, or monitor these agentic systems.
Paywall Vulnerabilities
The vulnerability extends beyond simple crawler detection. Many publishers, including National Geographic and the Philadelphia Inquirer, rely on client-side overlay paywalls where text loads on the page but remains hidden behind a subscription prompt.
While this content is invisible to humans viewing the page normally, AI agents like Atlas and Comet can parse the underlying code and extract the text directly.
In contrast, outlets like the Wall Street Journal and Bloomberg employ server-side paywalls that prevent the full text from reaching the browser until credentials are verified.
However, once a user is logged in, AI browsers can read and interact with articles on their behalf.
The problem intensifies when AI agents encounter blocked content. Research has documented that Atlas and similar systems employ sophisticated workarounds to reconstruct paywalled articles.
When prompted to summarize content from PCMag—whose parent company Ziff Davis sued OpenAI for copyright infringement—Atlas produced a composite summary by drawing on tweets, syndicated versions, citations in other outlets, and related coverage across the web.
This technique, described as reverse-engineering through “digital breadcrumbs,” allows AI agents to circumvent direct access blocks by assembling information from multiple sources.
Legal and Ethical Implications
OpenAI states that by default, it does not train its large language models on content users encounter in Atlas unless they opt into “browser memories,” though pages that have blocked OpenAI’s scraper will not be used for training.
Despite these assurances, ambiguity remains regarding how much data OpenAI extracts from paywalled content that users unlock for agents to read.
The situation highlights a critical gap: traditional defenses such as paywalls and crawler blockers are no longer sufficient to prevent AI systems from accessing and repurposing news articles without publisher consent.
As AI browsers continue to evolve and potentially reshape how users consume digital content, publishers face a challenging landscape.
Whether or not these tools achieve widespread adoption, the capability to bypass content restrictions fundamentally alters the relationship between AI systems and digital publishers.
If agentic systems represent the future of news consumption, publishers will require greater visibility into and control over how and when their content is accessed, used, and potentially repurposed by increasingly sophisticated AI agents.
Follow us on Google News, LinkedIn, and X to Get Instant Updates and Set GBH as a Preferred Source in Google.