Suppose you are on a bike trip or car whatever you prefer and to have some lunch you search for a restaurant on Google. You pull up the reviews and there are 200 of them, all five stars, half posted in the same week.
Every single review reads like it was written by someone who has never actually eaten food before. "Great ambience! Very tasty. The staff was helpful. Will visit again."
You close the tab and go anyway because what
choice do you have?
I think most people have quietly stopped
trusting what they read online. They just have not said it out loud yet.
When the Internet Was Still Honest
The early internet was a mess. GeoCities pages with tiled backgrounds and blinking text. Forums where people argued about nothing for pages. Amazon reviews written in 2003 by someone who bought a blender and had genuine feelings about it.
Nobody was optimizing anything. There was no
algorithm watching, no monetization waiting at the end, no brand partnership to
protect. People wrote because they had figured something out and wanted to say
so, or because they were bored, or because they were lonely.
The content was also terrible. It was badly
spelled, not even organized, and weirdly personal. But it was true.
Someone actually thought of that thing. Someone actually bought that blender.
That is the thing worth missing - the intent behind it.
The Moment Money Entered the Room
The shift did not happen all at once. It
crept in through money, slowly enough that nobody really noticed until it was
already done.
SEO Changed What Writing Was For
SEO arrived and suddenly the words on a page
were not just words anymore. They were signals to a machine that decided who
got seen. Write the right phrases enough times and traffic would come. Traffic
meant ad revenue. Ad revenue meant you could do this full time. And the moment
content became a business, it started behaving like one.
Content farms showed up around 2008, 2009.
Sites churning out five hundred articles a day on every conceivable topic,
written by anyone who would take two dollars per piece.
The writing was hollow but it ranked. Then
affiliate marketing layered on top and product reviews became sales pitches
wearing the costume of reviews.
Social Media Made Fakeness a Skill
Social media made the whole thing strange. Authenticity became a brand strategy, which is probably the most efficient way to destroy authenticity that anyone has ever invented.
You could watch people in real time calculating which version of being real would perform better with their audience. Instagram is full of lives that I think do not even exist.
People are wearing filters on social media. The platforms that people used to trust a lot are now battling with AI-generated content and bots.
The Platforms We Actually Trusted
Reddit, Google Reviews, Quora, these were the
places people went when they wanted something real. A genuine recommendation or
an honest opinion from someone with no reason to lie. Each of them, in their
own way, had a period where that promise held up.
1. Reddit
Reddit had genuine value for a while and it
came from friction. Communities had culture and new accounts were treated with
suspicion. To have your recommendation taken seriously you had to actually be
part of a community, which meant time, which meant a real person had to care
enough to spend that time.
Then it became too valuable to leave alone. Now there are threads in supposedly trustworthy communities that are quietly seeded. The "I tested twelve different mattresses and here is the one I kept" post with a six month old account and a suspicious posting history.
Entire recommendation threads written in the
same voice because they came from the same operation.
2. Amazon
Amazon reviews were the original trusted source before Google Reviews even existed. People built purchasing habits around them for years.
Now it has verified purchase badges that mean
nothing because review traders buy the product first specifically to earn the
badge. Entire seller networks that exist only to cross-review each other. The
five star system on Amazon at this point is closer to a lottery than a rating.
3. Quora and the Expert Problem
Quora launched with a genuinely different idea. Not random people answering questions, but the actual person who worked at that company, built that product, lived through that experience.
Early Quora had answers from people with real
firsthand knowledge and the quality showed.
Then the platform grew, the experts thinned out relative to the volume of questions, and the gap filled with people writing in the tone of expertise without the substance of it. Long confident answers about topics they had no real connection to.
Quora then introduced a Partner Program that paid people per question and per answer view, which turned what remained of a knowledge platform into a straightforward content farm.
People asked thousands of low quality
questions designed for traffic and answered them with content designed to rank.
Now a significant portion of answers are AI generated text dressed up with a profile
photo and a bio that says writer, thinker, traveler.
4. Yelp and Trustpilot
Yelp tried harder than most. It built an
aggressive spam filter, which created a different problem where legitimate
reviews got suppressed and businesses accused the platform of holding ratings
hostage unless they advertised. It got gamed from both directions
simultaneously.
Trustpilot positioned itself as the credible
alternative, the serious platform for business accountability. It is now
arguably worse than Google Reviews, partly because companies pay Trustpilot
directly for reputation management tools, which is a genuinely strange
arrangement for a platform whose entire value is supposed to come from
independence.
AI Held the Funeral and Wrote the Eulogy Too
If the previous era required hiring people to write fake content, this era requires almost nothing. A few prompts and you can flood any corner of the internet with text that is grammatically clean and completely empty.
Blog posts, reviews, comments, forum answers,
all of them generated in an afternoon by someone who has never used the
product, visited the place, or held an opinion about anything.
The part most writing on this topic skips over is that readers built the demand for this. Engagement data showed for years that people click listicles, share outrage, and scroll past anything requiring more than thirty seconds of attention.
Content farms were not irrational. They
looked at what was getting clicks and gave people more of it. AI content is the
final optimization of a signal that humans sent, with their attention,
consistently, over two decades, until the internet took it as instruction.
When Content Stopped Needing a Reader
This is where the story gets stranger. For
most of internet history, fake content was still written for humans to read.
The goal was to deceive a person, get a click, and make a sale. The human was
still the endpoint.
This article is a prime example of this. I am
writing this blog to promote my AI development services. Although
that is my secondary reason and the first one is that I actually like to write
and share my opinions.
But in most cases that is no longer true. AI systems now crawl the web, ingest published content, and use it to generate answers in AI Mode, in chatbots, in search summaries.
If the content they are reading was itself AI generated and factually hollow, those answers inherit the same problems and then deliver them with the confidence of a cited source.
The models themselves were trained on web data that included years of content farm output, SEO fluff, fake reviews, and hollow affiliate articles.
The model has no way to distinguish a genuine
forum post from someone who actually knew something from a 2019 article written
entirely to rank for a keyword. It consumed both equally.
What this means is that the web is now
producing content that is designed to be read by machines, repeated by
machines, and occasionally surfaced to humans who have no way of knowing where
it originally came from. The audience quietly changed and most people did not
notice.
What is Still Out There
Some things are still worth finding. A mechanic's YouTube channel where you can see the actual car and the actual problem. A food blog where the writer made the recipe badly the first time and wrote about that too, including the part where it stuck to the pan.
A Reddit comment from someone clearly annoyed
that the popular answer in the thread was wrong, written by a person who has
been living with the problem for years.
They exist. They are just buried under the
volume of everything else and finding them takes more effort than it used to.
The texture of real experience is still recognizable when you come across it.
Most people can feel the difference even if they cannot explain how they know.
The Mirror Has Not Lied to Us
The internet was always a reflection of what we paid attention to, what we rewarded, what we clicked on at eleven at night when nobody was watching.
AI content finished building what we started,
with better tools and lower costs, but toward the same destination we were
already heading.