For twenty years the hard part of being online was making the content. That part is now nearly free. The new hard part is being believed.
For about two decades, the limiting factor in digital strategy was production. If you wanted to show up in search, fill a feed, or run an email program, someone had to write the words, shoot the photos, and build the pages. Whoever could produce more, faster, usually won.
That economics broke. The cost of making content has fallen close to zero, and the result is a flood. Ahrefs analyzed about 900,000 new web pages in April 2025 and found that 74.2% contained some AI-generated text. Most of it is blended human and machine work rather than pure automation, but the direction is clear: producing content is no longer the bottleneck. A Stanford-led study published in early 2026 estimated that roughly 35% of newly created websites are now AI-generated.
The volume is not limited to articles. By April 2025, researchers from Columbia, the University of Chicago, and Barracuda found that about 51% of spam email was being written by AI. And for the first time on record, machines now make up the majority of the internet itself. Imperva's 2026 report found that bots accounted for 53% of all web traffic in 2025, with humans down to 47%.
So the internet got bigger. And less of it is human.
When content got cheap, trust got scarce
When something becomes infinitely abundant, it stops being valuable. That happened to content, and it pulled trust down with it.
People can feel the flood even when they cannot name it. A June 2025 Pew survey found that 76% of Americans say it is important to be able to tell whether something was made by a person or by AI, but 53% are not confident they can. That gap, wanting to know and not being able to tell, is the whole story in one line.
The skepticism is widespread. The Reuters Institute reported in 2024 that 59% of people worldwide, and 72% in the United States, worry about telling real from fake online. The 2026 Edelman Trust Barometer named misinformation and the spread of generative AI among the top forces eroding trust over the last five years, and showed people pulling their confidence inward, toward neighbors, family, and known local sources, and away from distant institutions.
For a business, that is the headline. Trust used to be granted by default to anything that ranked well or looked professional. Now it has to be earned at every touchpoint, because the default assumption has quietly flipped to "this might be fake."
Why everything started to sound the same
Underneath the flood is a quieter problem: the content is not just abundant, it is samey.
Language models work by predicting the most likely next word. Run that at scale across millions of sites and you get convergence. A Cornell study in 2025 found that AI writing assistance pulled people from different cultures toward the same generic, Western phrasing, flattening the specific detail that made each voice distinct. A 2024 study in Science Advances put a finer point on it: AI can make any one person's writing a little better while making everyone's writing more alike. The individual gains. The collective gets blander.
You have probably felt this without a study telling you. It is why a local plumber's website can suddenly read like a software startup, full of "seamless solutions" and "empowering your home." The model learned from corporate and SaaS marketing, so that is the voice it reaches for. The trouble is that customers notice the mismatch. When the words on a roofing company's page sound nothing like the way its real customers talk in its reviews, that gap reads as a signal: this is a template, not a person.
The cheapest thing to fake is now expertise
In late 2023, Sports Illustrated was caught publishing product articles under authors who did not exist, complete with AI-generated headshots and invented biographies. The fallout reached the top of the company. Since then the fakery has industrialized, and the platforms and regulators have started swinging back.
The FTC's rule banning fake and AI-generated reviews took effect in October 2024, and in December 2025 the agency sent warning letters to ten companies, with penalties that can reach about $53,000 per violation. Yelp's 2025 report is even more pointed for local businesses: it filtered out nearly 500,000 suspected AI-generated reviews and rejected more than 50,000 new business pages for spam, concentrated in exactly the categories where customers are most vulnerable, locksmiths, plumbers, roadside assistance, and garage door repair. Faking authenticity is now both a legal liability and a fast way to get removed.
What actually earns trust now
Here is the part that should encourage you, especially if you run a real business with real people.
The same forces flooding the web have made specific human detail more valuable than it has ever been, because it is the one thing that cannot be mass-produced. Search engines worked this out first. Google does not penalize content for being made with AI. It penalizes "scaled content abuse," low-value pages built to game rankings. What it rewards is demonstrated experience: a named author with real credentials, first-hand knowledge, original detail. A 2026 Semrush analysis found that human-written content holds Google's number-one position about 80% of the time, against roughly 9% for purely AI-generated pages.
Customers are aligned with the algorithm. Getty's research found that 98% of consumers say authentic images and video are pivotal to trusting a brand, and nearly 90% want to know whether an image was made by AI. A real photo of your team, your truck, or your actual finished job does more work now than any stock image or AI render, precisely because it takes effort that a spammer will not spend.
When everything generic is free, the specific becomes the premium.
That is the advantage. The named author. The real face. The actual job in the actual neighborhood. The detail only you have.
This is not an argument against AI
None of this means AI is the enemy, and it would be dishonest to pretend otherwise. The tools are genuinely useful, and the businesses that do best will not be the ones that refuse them. They will be the ones that keep a human in charge and let the machine assist: drafting, structuring, handling volume, while a real person supplies the judgment, the verified facts, and the voice.
Google's own position is the cleanest summary of the whole situation. It does not care whether a person or a machine produced the work. It cares whether the work is useful and trustworthy.
So the strategy for this era is not complicated, even if it is not easy. The web got bigger and cheaper, and trust became the scarce resource. You earn it back the same way every time, with the specific, verifiable, human detail that nobody can fake at scale. Be a real person, doing real work, and show your work. In a sea of synthetic sameness, that is no longer just nice to have. It is the whole advantage.