{"id":35181,"date":"2026-11-02T08:35:56","date_gmt":"2026-11-02T13:35:56","guid":{"rendered":"https:\/\/ianchadwick.com\/blog\/?p=35181"},"modified":"2026-11-02T08:35:56","modified_gmt":"2026-11-02T13:35:56","slug":"is-the-internet-dead","status":"publish","type":"post","link":"https:\/\/ianchadwick.com\/blog\/is-the-internet-dead\/","title":{"rendered":"Is the Internet Dead?"},"content":{"rendered":"
<\/div>\n

<\/g><\/svg><\/i> \"Loading\"<\/p>\n

<\/div>\n
Are Humans in Virtual Zombie Territory Every Time We Post Online?<\/h5>\n

\"Dead<\/a>I<\/span>t’s called the Dead Internet Theory. It’s not really<\/em> a theory in the scientific sense of the word; it’s more of a conspiracy story<\/strong> or, at best, a vague hypothesis. It has its roots in what is actually happening online. Something we can all see \u2014 and have<\/em> seen and experienced. Simply open Facebook or YouTube anytime to see it. In late 2025, OpenAI’s CEO, Sam Altman, posted on Xitter that: “i never took the dead internet theory that seriously but it seems like there are really a lot of LLM-run twitter accounts now.” He didn’t, of course, take responsibility for any of it.<\/p>\n

And it’s not really<\/em> dead because the zombies (bots and AI) have not fully<\/em> taken over. Yet. There is still a lot of human-generated content online (you’re reading some right now. But, I wonder, is it on its way to becoming a graveyard where human content is struggling to survive at the fringes while bots and AI spam control the majority of its real estate?<\/p>\n

The SpaceNazi Elon Musk’s rightwing swamp-platform, Xitter, cancelled more than 42,000 bot accounts using chatbots in July<\/a> alone, out of 1.7 million bot accounts it has shut down this year alone. And those didn’t include any of the “Bot accounts that label themselves as bots, post automated weather updates or stock prices, or surface information without pretending to be human…” These are still within Xitter’s allowable accounts policy<\/a>, and are not always neutral. All those cancellations didn’t make Xitter any less toxic, however: it’s still a rightwing swamp.*<\/strong><\/span><\/p>\n

Meta estimates<\/a> that “around 4-5% of Facebook accounts are fake, which works out to about 120-150 million accounts on a 3 billion user base. Independent research pushes that number higher, with many studies estimating 10-16% of Facebook accounts may be bots or fake<\/strong>, and ad traffic estimates reaching 50-70% bot activity.” (emphasis added) Meta has its own AI, of course, deciding what users get to see, who or what populates their feed, and what ads dominate.<\/p>\n

The Guardian<\/a> reported late in 2025 that “More than 20% of videos shown to new YouTube users are \u2018AI slop\u2019, study finds.”
\nIt added, “Low-quality AI-generated content is now saturating social media \u2013 and generating about $117m a year, data shows.”<\/p>\n

The video-editing company Kapwing surveyed 15,000 of the world\u2019s most popular YouTube channels \u2013 the top 100 in every country \u2013 and found that 278 of them contain only AI slop.
\nTogether, these AI slop channels have amassed more than 63bn views and 221 million subscribers, generating about $117m (\u00a390m) in revenue each year, according to estimates…
\nA Guardian analysis this year found that nearly 10% of YouTube\u2019s fastest-growing channels were AI slop, racking up millions of views despite the platform\u2019s efforts to curb \u201cinauthentic content\u201d.<\/p><\/blockquote>\n

Canadian author Cory Doctorow coined the term Enshittification<\/strong><\/em><\/a> for what has happened to the internet. Wikipedia tells us<\/a> it’s also known as “platform decay … a process in which two-sided online products and services decline in quality over time.” The Dead Internet Theory<\/strong> is more about takeover than decay, however. It’s the result of tech companies pushing more and more AI-generated content online, encouraging more and more users to create content with their tools, all the time trying to justify their immense investments in their corporate facilities, data centres, and programming efforts by making use of AI so ubiquitous it feels almost mandatory. And yet, doing it all without moral or ethical concerns about the results.<\/p>\n

But again I ask: is the internet really<\/em> dead? Or just being buried alive? Do the data match what we seem to see online every day?<\/p>\n

As Magnexis Development Team concludes in its case study<\/a>, “The Dead Internet Theory is best understood as a mixture of accurate observations and unsupported extrapolations.”\u00a0But what does that mean to the billions of users who are online every single day, who experience the growing amount of AI-generated content, whose social media feeds are increasingly filled with what has been nicknamed “AI slop”? The report continues:<\/p>\n

The strongest evidence supports the idea that automation, AI-generated content, bots, and algorithmic curation now play a far larger role than they did a decade ago. However, there is no credible evidence that most online activity is secretly generated by AI or that human participation has been broadly replaced.
\nRather than a “dead” internet, today’s web is more accurately described as an AI-augmented internet\u2014one where human and machine-generated content increasingly coexist, creating new challenges for authenticity, trust, and information quality.<\/p><\/blockquote>\n

“AI-augmented” doesn’t sound a lot more friendly than just dead. Machine translation, for example, represents another AI usage that seldom gets much attention. But a team of Amazon Web Services researchers<\/a> found that “roughly 57% of all web-based text has been AI-generated or translated through an AI algorithm.” And then you have AI translating AI. As noted in a Forbes magazine<\/a> article from 2024:<\/p>\n

[W]hen generative AI software relies solely on content produced by genAI, the responses begin to degrade, according to the study published in Nature last month… If human-generated data on the internet is quickly being papered over with AI-generated content and the findings of Shumailov\u2019s study are true, it\u2019s possible that AI is killing itself\u2014and the internet.<\/p><\/blockquote>\n

How much online content is really created by AI? Is it as much as we imagine or believe we’re seeing? In 2023, some AI experts predicted AI-generated online content could reach as high as 90% of the total by 2025<\/a>. Clearly, that has not happened. But by May, 2026, a Graphite Research study<\/a> found “AI Now Writes as Many Online Articles as Humans.” It found the share of AI to human-generated content was roughly equal:<\/p>\n

ChatGPT launched in November 2022. Within the first 12 months, the percentage of primarily AI-generated articles jumped to 36%, and reached 48% by 24 months.
\nHowever, since Q1 2025 the percentage of primarily AI-generated articles has plateaued at roughly 50%.<\/p><\/blockquote>\n

How does the average user, without any specialized tools or training, learn to differentiate human-made from AI-made content? There are a lot of websites that offer tips, advice, and suggestions for identifying it, but… are these sites AI-generated themselves? Even people trained to do so only get it right half the time<\/a> (emphasis added):<\/p>\n

Advances in generative AI technology have made it easier than ever for anyone to manufacture increasingly realistic synthetic media (colloquially known as deepfakes) at faster speeds, larger scales, and with more customization than ever. This in turn has led to synthetic media increasingly being used for harmful purposes, including disinformation campaigns, nonconsensual pornography, financial fraud, child sexual abuse and exploitation, and espionage. As of today, the principal defense to combat deceptive synthetic media depends in large part on the human observer\u2019s perceptual detection capabilities\u2014their ability to visually or auditorily identify AI-generated content when they encounter it. Yet the growing realism of synthetic media impedes this ability, heightening people\u2019s vulnerability to weaponized synthetic content. Moreover, people overestimate how capable they are at identifying synthetic media,<\/strong> further exacerbating the problem.<\/p><\/blockquote>\n

In July, 2026, Pew Research<\/a> surveyed 10,000 websites and found that “10% show significant signs of AI authorship.” However, when examining pages published after<\/em> ChatGPT was released, researchers found “signs of AI authorship can be found in over one-third of pages.” Ten percent seems small, but as the report notes, the trend toward using AI content rather than human-made content is growing rapidly.**<\/strong><\/span><\/p>\n

The Stacc<\/a> report compiled data from 50 sources to examine AI’s usage. In a study of 900,000 pages in April 2025, they found “74.2% of newly created web pages contain AI-generated content.” Their blog post shows other significant numbers from these data sources:<\/p>\n