Is AI content destroying the internet or transforming it?

AI content: Is it killing the internet?

Artificial intelligence has become one of the most influential forces shaping the internet today. The rise of tools like ChatGPT, Claude, and Perplexity is not only changing how people search for information but also how publishers, journalists, and reviewers can survive in a digital economy that once depended on traffic and advertising. What is now being called AI content has become both the fuel and the fire of an evolving information ecosystem, raising serious questions about sustainability, trust, and the future of journalism.

The rise of AI content as a primary source

For decades, search engines acted as the gateway to the internet. Users typed in queries, browsed links, and eventually landed on publishers’ websites where their attention could be monetised through advertising or subscriptions. AI tools have disrupted this model by removing the need for the middle step. Instead of directing users to links, AI systems deliver answers directly.

Google’s rollout of AI Overviews and AI Mode has accelerated this shift, often displaying summarised responses on the results page without crediting or sending traffic to original sources. For readers, this is convenient. They don’t need to sift through multiple pages to find an answer. For publishers, it is catastrophic. Their content is scraped, repackaged, and presented without compensation, effectively cutting them out of the value chain.

How publishers are losing visibility

Reports confirm what publishers have been feeling for years. A study by Enders Analysis, based on Sistrix data, found that news visibility in Google results has collapsed. The Mirror lost 80 percent of its presence since 2019, while The Mail lost more than half. Even subscription-based outlets like the Financial Times saw significant declines.

This isn’t a matter of poor strategy by publishers. It is structural. AI content intercepts audiences before they reach original sources, making the economic model that sustained journalism untenable. The once-dependable cycle of clicks, subscriptions, and advertising revenue is being eroded by AI-powered platforms that extract value without returning any.

Beyond news: The wider impact of AI content

Journalism is not the only industry under threat. Science, education, reference sites, and health information are being hit even harder. A chart published by The Economist shows health-related information as the most negatively impacted sector. Users who once sought advice from vetted medical organisations now receive AI summaries that may or may not be trustworthy.

This shift coincides with a measurable change in user behaviour. Apple reported its first-ever decline in Safari search volume earlier this year, citing users moving towards AI chat tools instead of traditional search engines. A TechRadar survey also found that 27 percent of US users and 13 percent of UK users now begin their research with AI assistants rather than search engines.

Why AI content risks eating itself

The long-term danger of AI-driven information is that if publishers cannot fund original reporting, fewer facts will be gathered in the first place. Good journalism requires reporters, editors, and fact-checkers roles that are expensive to sustain. If these vanish, AI tools will be left with only outdated articles, promotional press releases, or propaganda to summarise.

The result would be a hall of mirrors: summaries of summaries, hallucinations, and content loops with no original reporting to ground them. In other words, AI content could eventually eat itself, collapsing the very ecosystem that allows it to function.

The collapse of online reviews

The issue extends beyond journalism into the review ecosystem. Online reviews once provided consumers with trustworthy feedback, offering transparency that adverts could not. Independent reviewers on YouTube and blogs built loyal audiences through honesty. Amazon’s user review system gave everyday shoppers a voice.

However, manipulation soon followed. Fake reviews, incentivised ratings, and bot farms undermined trust. Now, AI tools summarise reviews without distinguishing between genuine feedback and manipulation. This further erodes incentives for trusted reviewers to produce high-quality work, as their efforts are scraped and repackaged without attribution or compensation.

Creators fight back against AI content scraping

Not all content creators are standing idly by. Musician Benn Jordan developed a tool called Poisonify, designed to stop AI models from training on music without consent. It works by adding imperceptible “adversarial noise” that corrupts scraped data while remaining unchanged to human ears. Similarly, media companies are experimenting with licensing deals, lawsuits, and technical barriers to block AI crawlers.

The New York Times has already sued OpenAI and Microsoft for training models on copyrighted journalism without permission. Others are pushing for new business models, such as Tollbit’s “paywall for bots”, which charges AI crawlers for accessing content, or ProRata’s system, which redistributes ad revenue generated from AI answers back to contributing publishers.

The business model behind AI content

Google’s dilemma illustrates the business challenge. Its primary goal is not to give users the best possible answer, but to serve an answer while keeping them engaged enough to click ads. With AI answers now appearing on top, links receive fewer clicks, undermining publishers.

Meanwhile, AI companies themselves are following the blitzscaling strategy of tech startups, growing rapidly, attracting investment, and disrupting industries without worrying about profitability. This makes AI content even more disruptive: even if these companies are unprofitable, they can still destabilise journalism, reviews, and education.

The problem of trust in AI content

Perhaps the most pressing issue is trust. AI systems are marketed as superintelligent, unbiased, and authoritative. In reality, they reflect the data they are trained on, the assumptions of developers, and the incentives of the companies behind them. Bias can be accidental or intentional, but either way, users may take the outputs at face value.

The danger is greatest in areas like health, science, and politics where expertise is hard-earned. Pew Research reports that trust in scientists has already declined steadily over the past five years. If AI tools become the primary interpreters of truth, public understanding risks being built on shaky foundations.

Authenticity and the battle for the future

In response, publishers and platforms are shifting towards personality-driven content. Readers and viewers are more likely to follow individual voices, journalists, YouTubers, Substack writers than faceless platforms. The Wall Street Journal even advertised for a “Talent Coach” to help journalists build personal brands.

Yet even this strategy faces threats. AI-generated influencers with synthetic faces and voices are already attracting followers. If authenticity is the last defensible moat, the next arms race may be in proving what is real.

Towards a sustainable future for AI content

The current disruption echoes earlier moments in media history. Napster nearly destroyed the music industry, but artistes adapted through streaming, touring, and direct fan support. Newspapers seemed doomed by the internet but reinvented themselves for the digital age. The future of AI content will likely follow a similar path: not extinction, but adaptation.

New revenue-sharing models, licensing deals, and technical safeguards will play a role. More importantly, the demand for truth, context, and accountability will remain strong. AI content may dominate distribution, but it cannot replace the human work of discovery and interpretation.

Why use a VPN

SECURITY: Our secure VPN sends your internet traffic through an encrypted VPN tunnel, so your passwords and confidential data stay safe, even over public or untrusted Internet connections.

PRIVACY: Keep your browsing history private. As a Swiss VPN provider, we do not log user activity or share data with third parties. Our anonymous VPN service enables Internet without surveillance.

FREEDOM: We created ProtonVPN to protect the journalists and activists who use ProtonMail. ProtonVPN breaks down the barriers of Internet censorship, allowing you to access any website or content.

Get ProtonVPN Now

Why the stakes are higher than we think

AI content is more than a technological trend. It is reshaping the economics of the internet, the viability of journalism, and the credibility of information. If unchecked, it risks hollowing out the ecosystem it depends on, leaving only recycled summaries and corporate PR.

But history suggests adaptation is possible. The internet will continue to evolve, and so will the ways we fund and value information. What matters now is ensuring that the foundations of truth and accountability survive long enough to support whatever comes next. AI may change the form of content, but the need for reliable information will never disappear.

________________________