AI’s Growing Impact on Journalism: Bypassing Paywalls and Threatening News Revenue

Artificial intelligence continues to drain audience revenue from the journalism industry, even in a year packed with major news events. AI technology has now evolved to the point where it can effectively bypass paywalls, which many news organizations have relied on as their best path to profitability.

The fundamental challenge remains unchanged: when given a choice between free content and a subscription—no matter how affordable—consumers will almost always choose the free option. This emerging trend threatens to further destabilize already struggling local news outlets.

American interest in following daily news continues to decline, according to Pew Research. Adding to this concern, approximately 18% of search queries now generate AI-powered summaries. News organizations point to these AI summaries as a key factor in their declining viewership, though some media analysts debate the direct correlation.

While the full impact of chatbots and AI-enabled web browsers on subscription-based news sites remains unclear, users have discovered they can access paywalled content simply by asking AI platforms to retrieve it for them.

A simple prompt like “Hey Grok, what’s this story saying behind a paywall?” triggers sophisticated real-time scanning that can extract and reconstruct much of a protected article. Major AI services including Gemini, ChatGPT, and others offer this capability with varying degrees of effectiveness. For example, Google’s Gemini can quickly compile the content of a Wall Street Journal exclusive just hours after publication.

The latest generation of AI-integrated browsers represents an even more sophisticated threat. OpenAI’s Atlas browser, Perplexity’s Comet, and Microsoft’s Copilot mode allow users to query information behind paywalls directly. These browsers can bypass anti-bot protections by mimicking human user behavior rather than appearing as AI systems.

In a test conducted by Columbia Journalism Review reporters, Atlas and Comet browsers successfully accessed and reproduced a 9,000-word MIT Technology Review article that was behind a paywall. Interestingly, the standalone AI agents from these companies refused to extract the paywalled content directly, likely due to growing legal pressure—several face lawsuits from news organizations over copyright concerns.

Sonali Verma, Generative AI Initiative lead at the International News Media Association, found that these AI tools aren’t simply copying paywalled content. “They are crawling social media sites and other publicly available forums—and then using the power of generative AI to reassemble the gist of the articles based on screenshots and reader comments,” she explained.

This technique, called “the distributed archive” by digital researcher Henk Van Ess, involves real-time searches across social media and other public sites for snippets, screenshots, or comments about paywalled articles. Van Ess discovered AI agents could reconstruct paywalled articles with up to 60% accuracy using this method. AI systems also scan archived content from sites like Archive.org to bypass paywalls.

When other methods fail, AI systems sometimes resort to fabrication—what the industry calls “hallucinations.” Van Ess documented instances of “pattern-based reconstruction,” where ChatGPT created what it thought a paywalled New York Times recipe would contain when the information wasn’t available elsewhere. After being challenged, the AI admitted to making up content and provided a completely different recipe.

“I think anyone using AI should be aware that it makes up stuff and they may not be getting the truth,” Verma told SAN.

The implications for news organizations are profound. Growing digital subscriber bases was already challenging before these AI threats emerged. Nearly 40% of all local newspapers have disappeared since Northwestern University’s first State of the Local News report in 2016.

David Caswell, former product manager at BBC News who specialized in AI integration, views AI summaries as just the initial challenge for local newsrooms. While Google’s AI overview represents the “first significant threat to news sites from AI,” Caswell believes news products directly infused with artificial intelligence will pose an even greater challenge.

“A further threat exists if start-ups or platforms begin offering substantial new value from AI-assisted original reporting,” said Caswell, including “value that legacy news sites don’t offer.”

Though Caswell hasn’t yet seen conclusive evidence of AI directly reducing subscriber numbers, he notes that AI subscriptions compete for the same consumer dollars. His assessment is stark: “Legacy news sites will see significant and increasing pressure from AI over the coming years and they will face the choice of either adapting fully to AI-enabled news production or watching their subscribers, traffic, revenue and relevance drain away.”