# AI Slop and the End of Knowledge > Source: > Published: 2026-06-30 19:04:26+00:00 ###### [Artificial Intelligence](/us/basics/artificial-intelligence) # AI Slop and the End of Knowledge ## How can we find reliable information behind AI slop and paywalls? Posted June 30, 2026 [ Reviewed by Margaret Foley ](/us/docs/editorial-process) ### Key points - The internet has often been a reliable place in which to store one's knowledge of the world. - Disinformation and AI slop have drowned reliable information under a flood of false information. - Effectively accessing knowledge is less about search and more about using critical thinking. I was an early fan of the internet. I was amazed that all the knowledge of the world was available to everyone, everywhere, all the time. With cell phones, we could carry all the knowledge of the world in our pockets. But things have changed. The knowledge is there. But it is drowning under a flood of disinformation and AI slop. AI systems may make it harder to find reliable information. ## Cognitive Offloading of Knowledge Google that. The ability to quickly and easily access information has been the boon of the internet age. But what if you don’t have to keep as much knowledge in your own head? The internet moved even more of our knowledge to external platforms rather than our internal brains. Why hold onto knowledge when we could simply search anytime for anything, when you carry all the knowledge of [ the world in your pocket](https://www.psychologytoday.com/us/blog/mental-mishaps/201403/the-world-in-my-pocket)? We all engage in cognitive offloading. We store our knowledge in places other than our brains. We take notes. We take pictures. We keep calendars. We keep financial records. I wrote recently about some of the [ benefits and risks of offloading memory](https://www.psychologytoday.com/us/blog/mental-mishaps/202606/the-cost-of-convenience-calendars-google-and-ai) to external records such as calendars, phones, and the internet. ## The Risk of False Information But one risk I didn’t mention was the reliability of information on the internet. If most of the knowledge you look up is reliable, then using the internet for storing knowledge is great. But what if the information on the internet isn’t reliable? I have written frequently about the spread of disinformation, both for Psychology Today and in my academic writing (here’s a list of links to [my writing on disinformation](https://wp.wwu.edu/cognition/disinformation-blog-posts/)). Disinformation can both lead people to believe false information and doubt the reliability of true information. ## AI Slop Beyond disinformation, we now have to worry about the use of AI for creating false information, generally called AI slop. The products of AI systems are being spread widely on the internet. There are estimates that 60 percent of TikTok videos and more than [20 percent of YouTube videos](https://www.kapwing.com/blog/ai-slop-report-the-global-rise-of-low-quality-ai-videos/) are AI-generated—that is, not real things, but things presented as if they are real (Tangerman, 2025). Should you worry about AI slop on TikTok? Since many people search for health information on TikTok, we should all be concerned about the lack of reliable information. Images on [social media](https://www.psychologytoday.com/us/basics/social-media) are often AI slop, and catching the fake images has gotten more difficult. You can’t just rely on there being too many fingers and teeth like you could three or four years ago. But it isn’t just social media. Summaries of scientific literature made by AI contain errors. Iris van Rooji (2025) described an interesting case in which the AI-generated definition of a relatively common term in cognitive psychology was frankly erroneous. The writers at Science checked the ability of ChatGPT to provide accurate summaries for their articles. They noted that even with precise instructions, the system made errors and often missed the most important aspects of new research (Eisenstadt, 2025). And if you ask these systems to find scientific research to address a question, they often make up citations. I’ve demonstrated this in class for students and found that over half of the described research papers don’t actually exist. There have also been published academic papers withdrawn because of nonexistent references and fabricated figures created by AI systems. It is becoming more difficult to find accurate information. Make sure you check in with reliable websites. And no, AI systems are not qualified to provide mental health counseling and may cause harm. ## Paywalls and Reliable Information Reliable information still exists on the internet. Oddly enough, Wikipedia, an open source of information, generally provides accurate information. At one time, teachers advised against using Wikipedia. But now many of us consider it a reasonable place to start learning about a topic, like the old, printed encyclopedias you can probably still find in a library. Unfortunately, much of the reliable information on the internet is behind paywalls. Do you want to read a scientific paper on a topic? You have to get through a paywall. For me, as a member of a few academic societies and with a university library, I can get through the paywalls and access scientific articles. Do you want reliable news? Another paywall. Less reliable information is freely available. But reliable information is behind paywalls. [Artificial Intelligence](https://www.psychologytoday.com/us/basics/artificial-intelligence)Essential Reads Is it possible to use these current AI large language models to find reliable information? That depends. If the question is clear and the answer is generally agreed upon, then the AI will give good information. But if the answer is in dispute or if new evidence means we have updated our understanding, then AI systems may provide incorrect information. Reliable information is updated by recent evidence. People are good at tracking and [ updating memories, ](https://www.psychologytoday.com/us/blog/mental-mishaps/202606/are-we-still-friends-the-importance-of-memory-updating)but AI systems struggle with this. You can work with AI systems to generate more information. Mike Caulfield (2025) has suggestions about asking a series of prompts to help you generate reliable and well-sourced information. But using the systems in this fashion generally is going to require a subscription, so the reliable versions of AI systems are often behind paywalls as well. ## Knowledge in an AI Slop World What does this mean for accessing knowledge and finding good information? At one time, you had to learn how to use a library to access reliable information. You needed subscriptions to newspapers or had to pay for cable. For a brief period, during the golden era of the internet, much of the information in the world was freely available. But now we must critically evaluate what we find on the internet to discern reality from disinformation and AI slop. References Caulfield, M. (2026). Four Tips for Better Prompting: A generalizable and durable approach to investigating an issue with an LLM. [https://mikecaulfield.substack.com/p/four-tips-for-better-prompting](https://mikecaulfield.substack.com/p/four-tips-for-better-prompting) Eisenstadt, A. (2025). Can ChatGPT help science writers? Science. [https://www.science.org/content/blog-post/can-chatgpt-help-science-writ…](https://www.science.org/content/blog-post/can-chatgpt-help-science-writers) Tangermann, V (2026). TikTok Has Been Completely Taken Over by AI Slop: It's practically inescapable. Futurism. [https://futurism.com/artificial-intelligence/tiktok-taken-over-ai-slop](https://futurism.com/artificial-intelligence/tiktok-taken-over-ai-slop) van Rooij, I. (2025). AI slop and the destruction of knowledge. [https://doi.org/10.5281/zenodo.16905560](https://doi.org/10.5281/zenodo.16905560)