Last week the Internet crossed a dubious record: AI-generated content now outnumbers human-written content on the internet. Research analyzing material published between 2020 and mid-2025 indicates that machine learning tools overtook human-written work late last year. This overwhelming volume means that generative AI content now makes up an estimated 57% of all online material. In text alone, analysis suggests that more than half of longer English-language posts on LinkedIn are AI-generated. You can imagine that a similar trend will also apply on images, videos, code, music and any other form of content that can be generated with AI.
This deluge of content is defined by the term “AI slop”. While the content can be informative and entertaining, it’s considered the advanced iteration of internet spam. It is characterised by low-quality text, videos, and images generated by AI. We are seeing platforms positively “sloshing” with this material, from strange AI-generated images of “Shrimp Jesus” on Facebook to absurd AI-generated cat soap-opera videos filling up feeds. This has been largely seen as a threat by creators, including MrBeast, who warned that AI generated Slop is threatening creators livelihood.
The Dead Internet Theory
The rapid spread of AI slop has severely blurred the lines between genuine and synthetic content, accelerating the degradation of online services. Cory Doctorow termed the gradual “enshittification” of the internet and the more extreme related Dead Internet theory.
We are quickly moving toward a reality where anything seen online might be AI-generated. The capacity of new tools like OpenAI’s Sora, which allows users to create highly engaging short AI video clips, means that users can be easily fooled by seemingly real content, such as security camera footage or deepfakes. Though Sora videos may be watermarked, sites already exist to remove those marks, ensuring that the world increasingly operates with “alternative facts”.
This flood of low-quality output has severe implications for content discovery. Interestingly, while AI content growth was dramatic following the launch of ChatGPT, data shows the proportion of AI-generated articles plateauing over the 12 months leading up to May 2025. A key hypothesis for this plateau is that AI-generated articles do not perform well in search. This suggests that services like Google may be having some success in suppressing the slop. Furthermore, some platforms are reacting directly to user backlash; Pinterest, for example, recently introduced controls allowing users to limit how much generative AI imagery they see in their feeds and Spotify has also removed 75 million AI generated tracks and is introducing new protection measures for artists who may suffer impersonation.
The flip side of the coin: anyone can be a creator
While the sheer volume of low-quality content causes widespread concern, the rise of powerful generative AI tools is also fundamentally changing the economics of content creation and this change is largely positive for aspiring creators.
AI is the latest step in a long lineage of technology that has served to blow open the floodgates of creativity and expression. Technology’s broad arc bends toward more affordable and accessible tooling, and AI accelerates this trend. What once required specialized knowledge and expensive equipment is becoming accessible to everyone.
Gen AI is expected to be net-net good for movie-making by making it much cheaper and faster to create high-quality material. Ultimately, Gen AI will be beneficial for anyone with creative DNA. Although most of the resulting output will be low-quality garbage, a significant amount “won’t be garbage,” leading to the emergence and anointing of a new generation of creatives.
For better or worse, early signs indicate social users find AI-made videos highly engaging. Sora hit No. 1 in free apps in Apple’s App Store on Friday, even though you need an invitation to use it. And those cat-family videos? They have millions of views on Instagram.
The Financial Times (source)
Implications for the Creator Economy: Taste as the Last Scarce Resource
The primary challenge and opportunity for the creator economy and related startups is differentiation. When content production is industrialized, the value shifts.
In an era of ubiquitous AI slop, “Taste” is the last scarce resource. Taste is what separates substance from slop. The market is demonstrating a deep hunger for taste, as evidenced by positive reactions to elegant, thoughtful AI branding campaigns.
For creators and the platforms that support them, the strategy moving forward must align with the counterculture that inevitably emerges when mass production dominates.
New companies in this environment will emerge in three main phases:
- Phase I: The creation of the AI tools themselves and the underlying infrastructure.
- Phase II (The VC Sweet Spot): The ways we can organize, filter, and curate content.
- Phase III: Getting paid. As previously mentioned on VC Cafe, platforms that pay creators stand to benefit from adoption.
As more content exists, future platforms and services that help users sort through the slop to find the “good stuff” will be invaluable. The next wave of successful companies will be those that prioritize injecting taste into every fiber of their product, thereby separating substance from slop.
