Let us filter AI slop, you cowards
The Verge argues that platforms should let users filter AI-generated posts instead of only labeling them. Current labels and provenance systems are portrayed as too weak to solve the problem.
Intelligence analysis by GPT-5.4 Mini
The piece says AI labels on platforms like YouTube, Instagram, TikTok, Pinterest, and DeviantArt do not give users enough control over what they see. Its core argument is simple: if companies think their AI labeling works, they should let people turn AI content off.
The article says social apps are putting name tags on robot-made pictures, but that is not enough. It is like labeling spoiled food but still serving it. The writer wants a button that hides the robot stuff completely.
Analysis
The case for a filter
The article argues that labels for AI-generated content are not enough. YouTube, Instagram, TikTok, Meta, and Spotify have all expanded authentication or disclosure systems, but those systems still leave users stuck scrolling past AI-made material they may not want to see.
The author’s point is that a simple filter would be the honest test. If a platform really believes its labels and metadata detection work, it should let users choose an "AI" toggle and hide that content from feeds. In practice, the article says, that would force companies to confront how limited their current systems are.
What the examples show
The Verge points to DeviantArt and Pinterest as rough examples of existing controls. DeviantArt has an AI content setting, but it is buried in account menus and does not appear to make a dramatic difference. Pinterest has category-based AI content controls, but the article says they are also hidden and imperfect.
The article then argues that broader platforms would likely fare no better. It says provenance systems like C2PA and SynthID depend on metadata or watermarks that can be missing or removed, while detection systems are only partial solutions. The result, according to the piece, is a familiar pattern: companies present AI labeling as a fix, but users still lack a practical way to avoid AI slop in their feeds.
The article frames this as both a product issue and a credibility issue. If platforms say AI authenticity matters, the author argues, they should prove it by giving users control rather than just adding labels that may not change what gets shown.
Key points
- The Verge argues that AI labels should come with a real filter option.
- The article says current disclosure systems do not meaningfully let users avoid AI content.
- DeviantArt and Pinterest are presented as weak or hidden examples of existing controls.
- Provenance systems and detection tools are described as incomplete and easy to defeat.
- The piece treats filter buttons as a test of whether platforms believe their own AI-safety claims.
If platforms add real AI filters, users who dislike synthetic content could more easily curate their feeds. That could also pressure companies to make labeling and provenance systems more reliable, since the filter would show whether they actually work.
The article suggests the likely failure mode is that filters would be hard to find and easy to bypass, so they would mostly look good on paper. If provenance data is missing or stripped away, platforms may still keep serving AI content while claiming they have addressed the issue.
