Faceless AI-generated YouTube channels have recently emerged as an issue. In response, one tech YouTuber named f4mi has devised a way to disrupt the AI bots that scrape content from her videos.

In her latest video (via Ars Technica), she showcased a method of adding gibberish to subtitle files to confuse AI tools. This method allows her to inject misleading data into the AI training processes.

Quote: “All this AI slop that’s appearing on every social media platform is not made by robots trying to steal our jobs, it’s made by humans trying to make money using AI to launder other people’s work.”

This innovative approach echoes methods previously seen where users trick AI by inserting positive keywords invisibly in their resumes.

Background: Recently, f4mi faced issues where subtitle files were too heavy, causing playback problems. The .ass format she used is typically for more customizable subtitles and can store large quantities of hidden text.

As an experiment, she successfully illustrated how AI models could be misled, creating an inconvenience for those attempting to extract content from her videos without proper attribution. This strategy aims to make it harder for those using AI to replicate her work while reminding viewers of the ongoing struggle against content theft promoted by automated systems.