CHANGE (IT BACK): Tech companies are fighting the AI “slop” they helped create.

It appears in everything from fake biographies and questionable recipes to music, videos, social media posts and online reviews.

In response, technology companies are beginning to remove spam and low-quality content, label AI-generated material, and give users more control over how much AI content they see. Companies such as LinkedIn, YouTube, TikTok and Substack have introduced different methods for identifying, removing or limiting this material.

The problem is complicated because many of the same companies trying to control AI slop also created or promoted the tools that make it possible. AI-generated content can increase engagement, but excessive amounts of low-quality material can eventually become a liability because users have difficulty finding the content they want.

But the real problem might be that as AI-generated slop increasingly pollutes the internet, newer models risk being trained on the slop generated by older models — a phenomenon known as Model Collapse. It’s in Big Tech’s interest to stop that from happening.