AI is reading popular books on YouTube for free, and publishers don’t know how to stop it

This is not an isolated case. YouTube has quietly become the largest free audio library in the world – and most of it is pirated copies.

How does it even go?

YouTube has a Content ID system that automatically recognizes protected content – but that system was originally built for music. It almost doesn’t work for audiobooks. It is enough to slightly change the tempo, tone of voice or add background music and the system does not recognize the copy. Publishers have to manually submit each video separately, which is practically impossible for an industry that churns out hundreds of new titles every day.

AI has taken the problem to a new level. With tools like ElevenLabs or similar services, someone can turn an entire e-book into an audio recording with a convincing synthetic voice in a few hours, upload it to YouTube and turn on monetization. Pirated versions of new titles often appear on the platform hours after the official release.

The scale of the problem is serious

Research shows that as many as 35 percent of users who listen to audiobooks do so on YouTube – and that most of them access illegal copies. The audiobook industry is worth more than two billion dollars a year and is one of the fastest growing in digital publishing. How much of that was lost to piracy is difficult to estimate.

The Association of American Publishers recently partnered with the company Vermillio to automate the detection and reporting of pirated content, but YouTube confirmed in a press statement that it cannot pre-screen every video uploaded to the platform.

Author Shaun Rein discovered a fake version of himself on YouTube reading chapters of his own book – the voice was cloned from his public interviews. Stephen Fry was not lucky either – recordings of him and Harry Potter were used as early as 2023 to make an illegal clone of his voice.