Your collective defense against AI-generated spam and content farms
We made it our mission to prevent the web from becoming a useless and harmful place. That’s why today, Kagi Search introduces the first community-driven system to detect and downrank misleading AI-generated text, images, and videos inside search results.
It’s 2025, and the Internet we used to love is drowning in AI-generated noise. Content farms using AI for profit are manipulating search results in this race for the attention economy.
This makes us wonder: Who are we building the web for?
What is AI “slop” and how can we prevent it?
AI slop is misleading or low-value AI-generated content designed to manipulate rankings or attention rather than help the reader.
As per our AI integration philosophy, we are not against AI tools enhancing human creativity. But when it includes fake reviews, fabricated expertise, misinformation, content farms designed solely for profit rather than value, and systems that try to replace genuine human insight and connection, we know it’s hurting us, and we take it upon ourselves to take action.
At Kagi our ethos is to keep humans in control.
We’ve been battling the AI slope ever since we introduced our AI-generated image filters a year ago. Since our founding, we have actively downranked content filled with ads and trackers that offer little or no value to our members, leaving you inspired and empowered to take control of your search experience.
Slopstop now tackles the broad spectrum of deceptive AI-generated content: videos, articles, domains, and everything in between. From now on, you will see a display showing real-time AI slope scores within search results. As our CEO Vlad says:
“We believe AI slop is an existential threat to the internet that should be for humans. This is the first step toward our ultimate goal: eliminating AI slop so you never see it again.”
This initiative will give you even more control over what you see online, elevating high-value, trusted information from misinformation, news websites, false narratives and content farms. We will improve the system by learning from your feedback and creating more automated elements.
All Kagi Search users can now flag low-quality AI content (“AI slop”) in web, image, and video search results. We will verify these reports using our own signals. If a domain primarily publishes AI-generated content, we will downrank it in Kagi Search and flag it as AI sloppy. If a page is AI-generated but the domain is mixed (mostly not AI), we will mark the page as AI-generated but will not downrank it.
For media results, images and videos confirmed as AI-generated, they will be labeled as such and automatically downranked on the results page. Users can also choose to filter out AI-generated media entirely.
Power Pair: Slopstop and Small Web
AI is developing so rapidly that it is becoming complex to detect, but not impossible. Not all AI-generated content is harmful and misleading, but if a domain is only in the business of broadcasting AI content, we consider it substandard.
In parallel with fighting AI-generated degradation, we are implementing solutions to whitelist and increase verified human creators online through our Small Web initiative. Every AI glitch we flag makes authentic human content more discoverable. We want to prioritize creators who truly add value to the Internet, no matter what device they use.
The Small Web represents everything that AI’s recklessness threatens: authentic human voices, genuine creativity, and content created for passion rather than profit. Together, Slopstop and Small Web create a powerful defense against the commercialization and artificial pollution of the Internet.
Building the largest AI slope dataset to fight LLM hallucinations
In our quest Slopstop is a step towards an enhanced, trusted experience in the Kagi ecosystem. As a result of this initiative, we aim to create the largest dataset of AI-sloped domains on the web using in-house-built detection and a carefully curated community reporting system. In short, we are using AI to destroy the AI slope.
We will use this dataset to build our own AI content recognition technology to be used in our products as additional protection against AI-generated hallucinations, false claims, and misinformation, which we now know accounts for 30-41% of the failed response rates in most other chatbots.
Access to the database will be shared soon, if you want to receive updates you can express interest here.
Join the fight: protect the quality of your searches
The fight for Internet authenticity cannot be won without your support. We’re starting with this crowdsourced effort to help learn and develop the ultimate, automated solution. Every piece of harmful AI-generated content you identify helps create a better, more trustworthy search experience for everyone.
See something that qualifies as AI generated? Here’s how to flag it:
- Click the shield icon next to any search result
- Select “Report as AI-generated”
- Our review team takes it from there
See our documentation to learn more about how slopestop works. As always, we rely heavily on user input for all of our products, so if you have feedback or suggestions, share them on our forums.