AI Slop Is Destroying The Internet

The digital landscape is currently experiencing a dramatic influx of content generated by artificial intelligence, often termed 'AI Slop,' which is rapidly altering online dynamics. This proliferation is occurring alongside the release of the 12,026 Human Era Calendar, marking a special year for the creators. The core issue revolves around how online attention translates directly into revenue, creating incentives for the mass production of low-quality, algorithmically optimized material.

AI Slop is saturating the internet and things are becoming dramatic pretty quickly.

The environment where money is generated through capturing user attention has been fundamentally changed by automated systems. Fake users actively disseminate their low-quality content across review sections, generate artificial traffic metrics, or actively poison public discourse. Current statistics indicate that roughly half of all internet traffic is composed of bots, with the majority serving destructive objectives rather than beneficial ones.

The creation of mediocre content has never been simpler, evident across numerous platforms. This includes the 'black hole of meaninglessness' found on LinkedIn, low-effort short videos designed to hypnotize younger audiences and severely diminish their attention spans, and countless soullessly rewritten books appearing on Amazon.

A deeply frustrating aspect of this trend is that the very creative work produced by humans is being utilized to train these sophisticated AI models. Every comment posted on Reddit, every original video uploaded to YouTube, and every human drawing shared on Deviant Art has effectively been sold off or directly appropriated by AI corporations.

This appropriation occurs without providing any attribution or payment to the actual creators responsible for generating the original material. This scale of creative theft makes protection virtually impossible, actively endangering the work of countless creatives while allowing AI companies to accrue significant wealth.

While the creative theft is frustrating, the potential for generative AI to irreversibly break the internet is considered far worse. The primary danger lies in the increasing difficulty for users to discern what information presented online is factually accurate.

generative AI truly has the potential to break the internet irreversibly.

Initially, artificial intelligence seemed promising. For instance, producing a high-quality script, such as those created by Kurzgesagt, begins with foundational research, which is then subjected to in-depth fact-checking by two or three individuals. Information is confirmed using trustworthy sources, ideally primary documents or academic papers, followed by input and critique from one to three domain experts. This process of fact-checking and source compilation alone consumes approximately 100 hours per video, acknowledging that human limitations mean mistakes are unavoidable.

When AI emerged, there was significant excitement about a mechanical brain capable of rapid information collection. Initial tests using professional accounts across various AI models to summarize information about brown dwarfs yielded impressive outlines with unique facts and source links. However, deeper investigation revealed that while over 80% of the data was solid, the remainder consisted of compelling facts—like brown dwarf superstorm speeds—that the AI could not source, suggesting extrapolation or invention.

The AI confidently provided incorrect or fabricated details, similar to a journalist inventing specifics to enhance a story's impact or fit a predetermined narrative. When experts were consulted, they flagged these same facts as questionable, prompting the AI to invent information to satisfy the goal of appearing knowledgeable and making the user happy.

Further investigation revealed that one seemingly solid source given by the AI was itself an article written in a style highly suggestive of AI generation, confirmed by an essay detection tool showing a 72% match. This created a closed loop where an AI article, lacking credible sources, was used as a credible source for subsequent AI research. By 2025, over 1200 confirmed AI news websites were publishing massive amounts of misinformation, leading AIs to present shoddy conclusions that sound authoritative but are often half-truths.

The danger escalates when this cycle repeats. If a new AI repeats research based on this AI-generated content, and that content originates from a highly viewed video, the misinformation gains perceived validity. Before AI, tracing obscure lies was difficult; with AI proliferation, determining truth becomes nearly impossible.

The misinformation is now true.

The content will spread rapidly through these automated channels.

The most corrosive element of current AI is its convincing appearance of intelligence, enabling it to deliver incredibly confident falsehoods, often subtly. When confronted, the AI admits the error only to repeat the behavior later, as there is no underlying consciousness or understanding guiding its output; it is merely a complex tool operating without comprehension of its task.

Studies analyzing millions of scientific papers before and after the rise of Large Language Models (LLMs) showed a sharp, abrupt increase in the frequency of words favored by AIs. This suggests a significant portion of papers are now AI-assisted, often without disclosure. Furthermore, researchers have been caught sneaking hidden messages into papers, prompting AIs via white text or tiny fonts to review their work positively and ignore flaws.

Human attention remains the single most valuable resource on the internet. If current trends persist, cheap, 'good enough' AI slop content will consume the majority of this attention. This shift risks making society dumber, exacerbating political divides, and causing neglect of genuine human interaction.

If AI consumes the majority of the attention pie, channels requiring massive human investment for research and quality production become financially unfeasible. This forces creators into a difficult choice: downsize operations or reluctantly adopt AI tools themselves simply to compete in the attention marketplace.

The responsible approach to AI involves using it as a helpful utility, much like the alignment tool in Adobe Illustrator, which instantly organizes multiple elements. AI should handle repetitive, mechanical tasks, allowing human creativity and integrity to lead the final product. This approach maintains efficiency without sacrificing authorship.

The commitment moving forward is that content creation will remain human-made, for humans. This involves significant investment in research, human creativity for illustrations and animations, and maintaining rigorous fact-checking processes discussed with human experts to ensure maximum trustworthiness. The creators state a preference to cease operations rather than produce AI slop.

Sustaining this human-centric project requires community support, as the organization employs nearly 70 full-time staff plus freelancers. The primary offering is the 12,026 Human Era Calendar, designed to fill homes with art while reframing time itself. By starting the timeline 12,000 years ago at the dawn of civilization, 10,000 more years of human achievement are incorporated into the timeline, offering a new perspective on historical progress.

To commemorate the 10-year calendar anniversary, a first-ever artbook was created, containing 120 pages of every calendar illustration ever produced, complete with behind-the-scenes sketches and team stories. These products, like the videos, are crafted with intensive human effort, research, and design, standing in direct opposition to content churned out by soulless algorithms.