Dead internet theory: what's true, what isn't, and what it means for your website

Dead internet theory is the idea that most of what you see online is no longer made by real people. It says bots and AI now produce most posts, comments and articles, and the strongest version claims this was planned, often by governments or big companies, to shape public opinion. The planned, coordinated part has no evidence behind it. But some of the underlying trends are real and measurable: more than half of web traffic is now automated, a large share of new articles are mostly AI-written, and AI answers in search are reducing clicks to websites. For business owners, the practical takeaway is not panic. It is to clean up your data, protect your forms and give people clear proof that a real business with real experts stands behind your site.

What dead internet theory claims

The theory comes in a mild version and a strong version, and it helps to keep them apart.

The mild version is an observation. A lot of online activity feels fake. Replies look copied. Social feeds fill up with odd, repetitive posts. Search results seem crowded with thin articles that all say the same thing. People sense that they are talking to machines more often than they used to.

The strong version is a conspiracy theory. It says the internet as people knew it "died" around 2016 or 2017, that most users and content since then are bots or AI, and that this was done on purpose. In the most famous telling, the U.S. government is behind it, using AI to manipulate the world's population.

The mild version is worth taking seriously. The strong version is not supported by evidence.

Where the theory came from

Ideas like this floated around imageboards for years, but the version most people know traces back to a specific post. In January 2021, a user called "IlluminatiPirate" started a thread titled "Dead Internet Theory: Most of the Internet is Fake" on a small forum called Agora Road's Macintosh Cafe. According to Futurism's coverage of the thread, the post claimed that "the U.S. government is engaging in an artificial intelligence powered gaslighting of the entire world population."

The idea reached a much wider audience in August 2021, when Kaitlyn Tiffany wrote about it for The Atlantic in a piece titled "Maybe You Missed It, but the Internet 'Died' Five Years Ago." The article treated the theory as wrong, while explaining why it struck a nerve: much of the internet really did feel more repetitive and automated than it once did. That framing, "wrong but feels true," is still a fair summary five years later.

What changed after 2021 was the arrival of tools like ChatGPT in late 2022. Suddenly anyone could produce endless text and images for almost nothing. Many people who had laughed at the theory started to wonder whether it was turning from fiction into a forecast.

Claim vs. evidence

Here is how the main parts of the theory hold up against current data.

The parts with real evidence

Bot traffic really is more than half

The most cited data comes from the annual Bad Bot Report by Imperva, a Thales company. The 2026 edition, published in April 2026, found that automated traffic made up more than 53% of all web traffic in 2025, up from 51% the year before. Human activity fell to about 47%.

Two caveats keep this in perspective. First, "automated" is not the same as "harmful." Google's crawler, Bing's crawler, monitoring services and price checkers are all bots, and many are useful. Second, this is a measure of traffic, meaning requests to websites, not a count of people. One bot can make millions of requests. So "more than half of traffic is bots" is true, while "more than half of the people online are bots" is not what the data says.

AI writing is now about half of new articles

Graphite, an SEO agency, sampled more than 55,000 English-language article URLs from the Common Crawl web archive and ran them through three AI detectors. Its May 2026 report estimated that 49.9% of articles published in the first quarter of 2026 were primarily AI-generated. The share climbed fast after ChatGPT launched, then stayed near 50% from early 2025 onward.

The same researchers note that, in separate work, they found AI-generated articles largely do not appear in Google or ChatGPT results. They also point out real limits: AI detectors are not perfect, the sample covers only English articles, and Graphite sells SEO services. So the honest reading is "a lot of new pages are AI-written, but many of them are not reaching readers," not "the web is now mostly machine-made."

Spam, fake reviews and engagement farming

Fake activity is old, but AI makes it cheaper. In August 2024, the U.S. Federal Trade Commission announced a final rule banning fake reviews and testimonials, including AI-generated ones. The rule also bans buying reviews that must say something positive or negative, hiding insider reviews, and buying or selling fake social media signals like bot followers or views when they are used to mislead.

You have probably seen engagement farming yourself: accounts that post odd, eye-catching AI images or bait questions to collect likes and shares, then use that reach to sell something or grow a page. People often call this flood of low-effort AI material "AI slop." It is real, it is annoying, and it is one reason the dead internet idea feels believable.

Search engines are pushing back on scaled AI content

Google has been clear that AI use alone is not the problem. In a February 2023 Search Central post, Google said its focus is "on the quality of content, rather than how content is produced," and that its systems aim to reward content showing E-E-A-T: experience, expertise, authoritativeness and trustworthiness.

What Google does treat as spam is mass production. Its spam policies define scaled content abuse as "when many pages are generated for the primary purpose of manipulating search rankings." Examples include using generative AI to create many pages "without adding value for users," scraping content to build pages, and stitching content together from other sites. In other words, the cheap content flood is exactly what Google says it is trying to filter out.

AI answers and AI crawlers are changing the traffic deal

This is the trend most likely to affect your business. A Pew Research Center study tracked the Google searches of 900 U.S. adults in March 2025. When an AI summary appeared, users clicked a regular search result on 8% of visits, compared with 15% when no summary appeared. They clicked a link inside the summary itself on just 1% of visits.

On the crawling side, Cloudflare's August 2025 analysis compared how often AI companies crawl websites with how often they send visitors back. The gap was large: in July 2025, Anthropic's crawlers made about 38,000 crawls for every referred visit, and OpenAI's about 1,100. Cloudflare also found that about 80% of AI crawling over the prior year was for training models.

None of this means the internet is dead. It does mean more of your content is being read by machines, summarized somewhere else, and visited by fewer people directly. We cover how to show up in those AI answers in our guide to answer engine optimization for websites.

The parts that don't hold up

The theory's conspiratorial claims are unsupported, and it is worth saying why.

• No coordinator is needed to explain the trends. Bots exist because scraping, spam and fraud make money. AI content exists because it is cheap to make. Search engines and platforms fight both because fake activity hurts their products. These are ordinary economic forces, not a secret plan.
• There is no "death date." The 2016 or 2017 timeline does not match any measurable change. The biggest jump in AI content came after late 2022, years after the supposed date.
• Real people are still the point. Bots that scrape, click and post usually exist to reach or fool humans. Fake reviews only matter because real buyers read them. A web that was truly empty of people would not be worth spamming.

The fair conclusion: the internet is noisier and more automated than it used to be, and that has real costs. But it is not fake, and it is not run by a hidden hand.

What this means for businesses and website owners

For a business, the useful question is not "is the internet dead?" It is "how much of my data is real, and do people trust that a real business is behind my site?"

Your analytics may include bots

Google Analytics 4 already excludes known bot traffic automatically, using Google's research and the IAB's International Spiders and Bots List. You cannot turn this off, and you cannot see how much was removed. But that list only catches known bots. Newer or disguised bots can still show up as sessions with near-zero engagement, odd locations, strange referral sources or sudden spikes. If your traffic jumped and your leads did not, bots are one of the first things to check.

Form spam and fake leads waste real money

Bots fill out contact forms, quote requests and newsletter signups. At best this clutters your inbox. At worst it pollutes your CRM, throws off your cost-per-lead numbers and sends your sales team chasing people who do not exist. If you pay for ads, fake conversions can also teach ad platforms to find more of the wrong traffic. Our guide to connecting website forms to email or a CRM covers how to route leads cleanly, and our website security checklist covers bot protection more broadly.

Trust is becoming the main advantage

When anyone can generate a thousand generic articles, a generic article is worth less. What is still scarce is proof of a real business: named people, real credentials, real photos, real customers and real results. This lines up with what Google says it rewards through E-E-A-T, and with what AI tools look for when choosing sources to cite.

Some ways that shows up on a site:

• Human authorship. Put a real name and short bio on articles, with relevant experience. A byline from someone who has done the work beats an anonymous "admin."
• First-hand expertise. Share what you have actually seen: real project examples, lessons from client work, photos you took, mistakes you made and fixed.
• Original data. Even small numbers count. "We reviewed 40 local service websites and 28 had no phone number above the fold" is something no AI can copy from elsewhere, because it did not exist before you published it.
• Real reviews. Ask real customers for reviews on Google and industry sites, and never buy or fake them. Fake reviews are now a federal compliance risk, not just a reputation risk.
• Consistent business details. Make sure your name, address, phone and website match everywhere. See why NAP+W consistency matters.

Build channels you own

If AI summaries and platform feeds are sending fewer people to your site, the answer is to depend on them less. An email list, a customer text list, repeat-customer programs and a recognizable brand that people search for by name all reach real humans without a middleman. A visitor who joins your list once can hear from you again without you winning a search ranking every time.

Practical checklist

Use this as a working list for your own site.

Clean up your analytics

1. In GA4, set up an internal traffic filter so your own team's visits are not counted.
2. Build a simple report or exploration that flags sessions with very short engagement, unusual countries or odd referral domains.
3. Compare traffic spikes against real outcomes like calls, booked jobs and sales, not just form fills.
4. Mark only meaningful actions as key events in Google Analytics 4, so bots that load pages do not look like conversions.

Protect your forms

1. Add a honeypot field: a hidden input that real people never see or fill in. If it has a value, the submission is almost certainly a bot.
2. Add a low-friction check like Cloudflare Turnstile, which Cloudflare describes as a "smart CAPTCHA alternative" that most visitors pass without solving a puzzle. It can be added to any site, even one not hosted behind Cloudflare.
3. Validate on the server, not just in the browser, and limit how many submissions one source can send in a short time.
4. Check new leads before they trigger expensive steps, like ad conversion tracking or sales calls.

Show you are real

1. Add a real About page with names, photos and credentials of the people who do the work.
2. Put author bios on articles and link them to a page about each author.
3. Show your address, service area and phone number clearly, and keep them consistent everywhere.
4. Display real reviews and link to where they live, such as your Google Business Profile.

Publish content worth citing

1. Cite your sources and link to the original data, like we have done in this article.
2. Publish original data, case studies or before-and-after results from your own work.
3. Write fewer, better pages instead of many thin ones. Mass-produced pages can fall under Google's scaled content abuse policy.
4. Add structured data so search engines and AI tools understand who you are. Our guide to llms.txt, schema and Open Graph walks through it.

Own your audience

1. Offer a reason to join your email list, such as a checklist, guide or seasonal reminder.
2. Send something useful on a steady schedule so subscribers stay engaged.
3. Track how many leads come from direct visits, branded searches and email, not just organic search.

The short version

Dead internet theory is wrong about the big stuff: there is no evidence of a coordinated plot, no 2016 death date, and no shortage of real people online. But it points at real changes. Bots now make up more than half of web traffic, roughly half of new articles are mostly AI-written, and AI answers are sending fewer clicks to websites. For business owners, that makes proof of real people, real expertise and real customers more valuable, not less. Filter your data, protect your forms, show who you are, publish things only you could write, and build an audience you can reach directly.

Want a website that clearly shows the real people and expertise behind your business? Talk to us.