This video essay by Kurzgesagt addresses the growing crisis of "AI slop"βlow-quality, mass-produced, AI-generated content that is saturating the internet. The creators argue that this phenomenon poses a significant threat to the digital ecosystem by devaluing human creativity, eroding trust, and poisoning the well of collective knowledge.
The video details the channel's own experiments using AI for research, revealing that while AI can efficiently gather information, it is often "confidently incorrect," fabricating details to create a more compelling narrative. This leads to a dangerous misinformation feedback loop, where AI-generated falsehoods are published, scraped by other AIs, and subsequently presented as credible facts, making it increasingly difficult to discern truth. The creators discuss how this issue extends to academic publishing and is used to spread false narratives.
In response, Kurzgesagt reaffirms its commitment to producing human-made content, pledging to use AI only as a tool for efficiency without compromising creative or journalistic integrity. The video concludes with a call to action, urging viewers to support human-led projects by purchasing their 12,026 Human Era calendar and a new companion art book, thereby helping to sustain high-effort, trustworthy content in the age of AI.
The Rise of AI Slop and the Fight for a Human Internet
0:12 The Problem of "AI Slop": The internet is being flooded with low-quality, AI-generated content ("slop"), from soulless articles and videos to rewritten books. Today, about half of all internet traffic consists of bots, many used for destructive purposes like spreading misinformation.
1:15 Creative Theft: A core issue is that AI models are trained on vast amounts of human-created content (art, videos, text) scraped from the internet, often without permission, attribution, or compensation for the original creators.
1:52 AI as a Flawed Research Tool: The Kurzgesagt team tested AI for video research. While initially promising, they found the AI was "confidently incorrect," inventing compelling but false information to make its output more interesting.
3:28 Fabricated Facts: In a test case about brown dwarfs, the team discovered the AI had invented specific "facts" (like the speed of super-storms) that were not supported by real sources. Experts consulted by the team confirmed these facts were likely fabricated.
5:42 The Misinformation Feedback Loop: The team later found a popular, AI-generated YouTube video that presented these same fabricated brown dwarf facts as truth. This illustrates the core danger: AI misinformation gets published, becomes a "source," and is then scraped by other AIs, cementing falsehoods as facts online.
6:56 Impact on Science: This trend is already affecting academia. Studies show a sharp increase in AI-favored terminology in scientific papers, and some researchers have been caught embedding hidden prompts in their work to trick AI reviewers.
8:02 Competing with AI Slop: In an attention-based economy, high-effort human content is at risk of being drowned out by cheap, "good enough" AI content, which could make channels like Kurzgesagt financially unfeasible.
8:30 A Commitment to Human Creativity: The channel pledges to remain "made by humans for humans." They will use AI as a simple tool for efficiency (like an alignment tool in design software) but will never replace human research, creativity, and integrity.
9:49 A Call for Support: To continue their mission of creating well-researched, human-made content, Kurzgesagt asks for audience support through the purchase of their products, primarily the 12,026 Human Era Calendar and a new 10-year anniversary art book.