The internet used to have too little consent. AI may have created the opposite problem: too much of it to trust.
For most of the internet’s history, creating content was constrained by time, money and human labour. A journalist needed hours to report and write a story. A company needed designers and copywriters to produce advertising. A YouTuber needed equipment, editing skills and time. Even a simple blog required someone to sit down and write it.
Generative AI has changed that equation.
Today, a single person can produce hundreds of articles, images, videos, product descriptions or social-media posts in a fraction of the time. The consequence is not simply cheaper content. It is a fundamental change in what is scarce online.
A new analysis from Pew Research Center provides a useful measure of the shift. After examining nearly 490,000 English-language webpages collected through Common Crawl, Pew found that 10% of webpages in its July 2026 sample showed significant signs of AI authorship. Among pages published after ChatGPT’s November 2022 launch, that figure reached 35%. Pew cautions that AI-detection systems are probabilistic and should not be treated as definitive judgments about individual pages, but the aggregate trend is striking.
The implication is bigger than the rise of “AI content.”
When content becomes abundant, content itself loses scarcity value.
What becomes scarce instead is the ability to attract attention, establish credibility and demonstrate that something came from an actual person with something real to say.
AI Is Solving The Production Problem
The economics of digital publishing are changing because AI is dramatically lowering the cost of production.
Adobe’s 2025 research found that 73% of Indian businesses surveyed reported greater volume and speed of content ideation and production after adopting generative AI, while 67% reported increased productivity and efficiency. The survey covered 345 executives and 841 consumers in India.
That distinction matters.
If producing an article, marketing image or product description takes substantially less time, companies have an incentive to produce more of them. The limiting factor moves away from the supply of content and toward the capacity of audiences to consume it.
Pew’s data shows what this looks like at the internet level. Signs of AI authorship among .com pages rose from 1.09% in early 2021 to 9.35% in January 2026 in its samples. For .org pages, the corresponding January 2026 figure was 4.59%, while .edu and .gov pages were around 1%.
The technology therefore creates an unusual economic situation: the marginal cost of creating another piece of content can approach zero, while the amount of human attention available to consume it does not.
That creates a bottleneck.
And bottlenecks are where economic value tends to migrate.
The Attention Market Cannot Expand At The Same Speed
There is no shortage of things to watch, read or listen to.
There is a shortage of time.
People still have roughly the same number of waking hours. They cannot read 500 articles simply because AI can produce them. They cannot watch 2,000 videos because video generation has become cheaper.
This makes attention increasingly valuable.
Google illustrates the scale of the underlying market. The company says Search handles more than 5 trillion searches annually, while its AI Overviews feature has expanded to more than 200 countries and territories and over 40 languages. In the US and India, Google reported that AI Overviews were driving more than a 10% increase in usage for the types of queries where the feature appears.
But AI is not simply adding more information to search. It is also changing the way people interact with information.
Instead of clicking through ten websites, users can increasingly ask an AI system to synthesize the answer.
That creates a second scarcity: being the source that gets selected and surfaced.
A million websites may contain information about a subject. Only a handful may receive the user’s attention or be referenced by the systems deciding what information the user sees.
The competition is therefore moving from:
Who can publish?
to:
Who gets noticed?
Search Is Becoming A FIlter, Not Just A Directory
The traditional web operated largely like a giant directory.
You searched for something, received links and decided where to go.
Generative AI increasingly inserts a layer of interpretation between the user and those links.
Google says AI Overviews are designed to provide answers while still featuring prominent links to relevant websites. At the same time, Google’s own data says overall organic click volume to websites has remained relatively stable year over year, while average click quality has increased.
The important issue for publishers is therefore not simply whether AI eliminates clicks.
It is which clicks survive.
If an AI system can answer a basic factual question immediately, generic explanatory content becomes easier to replace. A website saying what inflation means, for example, has less differentiation than a publication containing an original interview with a central banker, proprietary data or firsthand reporting.
This creates an uncomfortable incentive.
The internet may produce more information while simultaneously making generic information less valuable.
Original information becomes more important precisely because AI can reproduce everything else so easily.
Trust Becomes Harder When Everything Looks Plausible
The deeper problem is not that AI can produce bad content.
Humans have always produced bad content.
The problem is that AI can produce convincing content at enormous scale.
That changes the cost of deception and the cost of verification.
A fake product review can be generated in seconds. A synthetic photograph can look convincing without representing a real event. An AI-generated video can put words into someone’s mouth. A fabricated expert profile can be surrounded by apparently credible text.
The distinction between “false” and “plausible” therefore becomes increasingly important.
Research from the Reuters Institute illustrates the tension particularly clearly. Its 2025 Digital News Report found that people expected generative AI to make news cheaper to produce and more up to date, but also less transparent, less accurate and less trustworthy.
Across its 48-market sample, overall trust in news stood at 40%.
That creates a paradox.
AI can make information production more efficient while making consumers more cautious about the information they receive.
And caution has an economic cost.
When people cannot easily determine whether something is authentic, they have to spend more time verifying it or retreat toward sources whose reputations already do that work for them.
Reputation Starts Behaving Like Infrastructure
This is where established brands, journalists, creators and specialist communities gain an unusual advantage.
A reputation cannot be generated as cheaply as text.
A company can use AI to produce 10,000 product descriptions. It cannot instantly manufacture decades of customer experience.
A publication can automate summaries. It cannot instantly recreate a reporter’s network of sources.
A creator can use AI to accelerate editing. But the reason an audience follows that creator may be the accumulated trust built over years.
The Reuters Institute’s 2026 Digital News Report found that 46% of respondents globally receive some news from creators of any type, while 27% receive news from news-focused individual creators or influencers. Yet respondents also considered creators less trustworthy and less impartial than some of their other attributes, including authenticity and relatability.
That distinction is revealing.
People may not automatically trust creators more than institutions.
But they increasingly value the sense that there is a real person behind the content.
That becomes particularly important in an internet where the cost of producing polished material continues to fall.
Authenticity Is Becoming Harder To Fake And Therefore More Valuable
The word “authenticity” is often used loosely. In an AI-heavy internet, it needs a more concrete definition.
Authenticity can mean proven origin, firsthand experience, accountability or identifiable human involvement.
A restaurant review from someone who actually ate at the restaurant carries a different kind of information from 500 automatically generated reviews.
A video shot by someone present at an event provides something a synthetic reconstruction does not.
An investigative article containing named sources, documents and original interviews offers information that cannot simply be produced by asking an AI model to “write a comprehensive article.”
This is why provenance technology is becoming important.
The Coalition for Content Provenance and Authenticity, or C2PA, has developed standards for Content Credentials that can attach cryptographically bound information about a digital asset’s provenance. The objective is not to declare every piece of content “real,” but to provide information about how an asset was created or modified.
That is a fundamentally different approach to content moderation.
Instead of asking only:
“Is this fake?”
the internet can increasingly ask:
“Where did this come from, and what happened to it?”
Platforms Are Already Putting A Price On Inauthenticity
The economics are beginning to show up in platform rules.
YouTube changed the terminology of its monetization policy in July 2025, renaming it “repetitious content” policy as “inauthentic content”. YouTube says repetitive or mass-produced material that does not provide original, authentic value is ineligible for monetization. It specifically includes AI-generated material using generic or unoriginal templates when it appears to be mass-produced without original insight or perspective.
That is significant because it shows where the platform itself sees the problem.
AI is not automatically the enemy.
Interchangeability is.
A creator who uses AI to research, brainstorm, edit or produce visual effects can still add something distinctive. But if thousands of channels can generate essentially identical videos from the same template, the content becomes less valuable to viewers and advertisers.
The scarce asset is not the production technology.
It is the reason someone should choose one piece of content over another.
AI Could Make Human Creativity More Valuable But Only If It Is Distinctive
There is a common assumption that AI will make human creativity less valuable because machines can create.
The opposite could happen in certain markets.
When everyone has access to the same generation tools, the tool itself stops being a competitive advantage.
Imagine two businesses with access to the same AI model.
Both can generate advertising copy, create product images, produce social-media calendars, and summarize customer feedback.
The advantage therefore moves somewhere else: proprietary data, brand identity, distribution, customer relationships, original ideas and human judgement.
Adobe’s 2025 creator research found that 86% of surveyed creators globally were using creative generative AI, but 69% were concerned about their content being used to train AI without permission. The same research found that unreliable output quality and uncertainty about how models were trained remained barriers to adoption.
The technology is becoming widespread. That makes the technology itself less distinctive.
The Biggest Winners May Own The Filters
This is perhaps the most important economic consequence.
In an internet where content is scarce, companies compete to produce it. In an internet where content is abundant, companies compete to filter it.
- Search engines filter.
- Recommendation algorithms filter.
- Newsletter filter.
- Creators filter.
- Communities filter.
- AI assistants increasingly filter.
The value therefore shifts toward whoever can answer a deceptively difficult question:
“Out of everything available, what should this person see?”
This explains why trusted brands and personalities can retain influence even as production becomes automated.
The Reuters Institute found that trusted news brands and official sources remain among the places people most frequently turn to when checking whether something online is true or false.
Trust, in other words, can function as a shortcut.
If verification becomes expensive, people rely more heavily on signals that reduce the cost of deciding what to believe.
The Internet May Enter An “Authenticity Premium” Era
For years, digital businesses benefited from removing scarcity.
Music became nearly unlimited through streaming. Information became nearly unlimited through search. Images became nearly unlimited through social media.
Generative AI pushes that process further.
But abundance creates its own scarcity.
- When everyone can produce an image, a real photograph becomes more interesting.
- When everyone can write an article, original reporting becomes more valuable.
- When everyone can generate a polished video, a credible person speaking from firsthand experience can stand out.
- When everyone can build an audience with automated content, a community that actually cares about its members becomes harder to reproduce.
The economic value of the internet could therefore move through a new chain:
Content → Attention → Trust → Authenticity → Influence.
AI makes the first layer dramatically more abundant. The remaining layers are still constrained by human behaviour. And that is why the next phase of the internet may not be defined by who can create the most content.
It may be defined by who can give people a reason to believe, remember and care about what they are seeing.
The internet’s next scarce resource is not information.
It is credibility.
