Here\u2019s what you\u2019ll learn when you read this story:<\/p>\n
First surfacing in 2021, the \u201cdead internet theory\u201d is the idea that the future internet will be largely driven by autonomous bots rather than real-world humans, creating an endless content factory driven by the engagement of solely digital entities.<\/p>\n
Although first inspired by the rise of algorithmic ranking systems (which are themselves autonomous), the arrival of large language models like OpenAI\u2019s ChatGPT has made the problem even worse.<\/p>\n
Report after report suggests that human-made content pre-AI is slowly dying due to link rot, and steadily proliferating bot content is taking its place, meaning the internet could be theoretically \u201cdead\u201d by the end of the decade.<\/p>\n
In early 2024, users\u2019 feeds across a variety of social media<\/a> platforms were bombarded with AI-generated images of various crustaceans (mostly shrimp) unsettlingly sporting the visage of Jesus. This self-described \u201cshrimp Jesus\u201d phenomenon, while mostly harmless on the surface (if not a little creepy for machine learning specialists and internet historians), represented something more sinister: It was a subtle glimpse into the emergence of a \u201cdead internet.\u201d<\/p>\n \u201cSome of these hyper-realistic images<\/a> have garnered more than 20,000 likes and comments. So what exactly is going on here?\u201d Jake Renzella and Vlada Rozova wrote about the phenomenon in The Conversation<\/a>. \u201cThe \u2018dead internet theory\u2019 has an explanation: AI and bot-generated content has surpassed the human-generated internet.\u201d<\/p>\n The \u201cdead internet theory,\u201d which was once a relatively fringe idea, has slowly garnered more and more attention, especially with the rise of AI-generated<\/a> content in the past few years. A 2024 \u201cBad Bot\u201d report<\/a> from cyber security firm Imperva estimated that nearly half of all traffic on the internet at the time was automated\u2014bots made up 42.3 percent of internet traffic in 2021, a number that jumped to 49.6 percent in 2023. Assuming that sort of growth continues continues, bots would be a clear supermajority by the late 2020s, just a few years from now.<\/p>\n Another report from the Pew Research Center<\/a> estimates that 38 percent of (mostly) human-made webpages from 2013 no longer exist, which is the result a process known as \u201clink rot.\u201d Because of these developments behind the world of online content, it\u2019s easy to draw the inevitable conclusion that, as Renzella and Rozova say, the internet is no longer \u201cfor humans, by humans.\u201d<\/p>\n In the case of a Shrimp Jesus, Renzella and Rozova mostly see just bot-supported engagement-farming at work. If you create an economic system where attention equals dollars (via ad revenue), then automating that brain-rotting process via bots is basically just an easy way to print money<\/a>. Taylor Lorenz, a former New York Times and Washington Post reporter and an overall expert on internet culture spotlights the rise of algorithmic content as the beginning of the end.<\/p>\n \u201cI think the internet was terminally ill before ChatGPT was announced and released,\u201d Lorenz says in a video<\/a> by YouTube science educator Kyle Hill. \u201cAlgorithmic ranking systems, which are AI driven, really set the stage for just endless, worthless pieces of content and for the whole internet<\/a> to be optimized in the most absurd ways.\u201d<\/p>\n Like almost every piece of tech ever made by humans, these techniques can be applied to more nefarious goals, such as supporting authoritarian regimes. A May 2025 review from the data analysis company NewGuard<\/a>\u2014designed to identify reliable sources of information\u2014found that there were more than one thousand news sites run almost entirely by bots. Some 167 of those sources were masquerading as Russian local news websites that published \u201cegregiously misleading claims about the Ukraine war<\/a> and primarily use AI to generate content.\u201d<\/p>\n Earlier this week, OpenAI CEO Sam Altman tweeted<\/a> that he never really took dead internet theory seriously, but that there are \u201ca lot of LLM-run twitter accounts now,\u201d seemingly unaware that his company is one of the chief architects of the modern internet\u2019s decaying state.<\/p>\n Of course, this doesn\u2019t mean that personal use of the internet\u2014messages among friends, private social media posts, and other correspondence\u2014will disappear. But with this growing decay<\/a>, users will need be more vigilant to discern what\u2019s real while also lamenting the internet that once was.<\/p>\n