Why Audiobooks Are Now a Major Piracy Target

Book piracy has always been a thing. Even in the early 17th century, famous authors, such as Miguel de Cervantes, were dealing with unauthorized copies and sequels of their work being published.
As soon as books could be (relatively) easily copied, there were those who did so without permission from the author or publisher.
As with most forms of piracy, the internet made it much easier to pirate books. It was, and still is, trivial to share unauthorized PDFs, ePubs, and text files online. The file sizes are small, and you can easily store many books in a small amount of space.
One such dataset, Books3, was a major part of the dataset used to train most AI systems. It represented over 12 percent of “The Pile”, a dataset used by most major AI companies.
Audiobooks have also been a major target for piracy. They are easily shared via MP3 or other compressed audio formats and, once again, have a relatively small file size that’s easy to store and share.
However, now there’s a new problem. YouTubers and streamers are creating their own audiobooks. What’s most interesting is that, while AI definitely is a major part of the problem, it’s not solely responsible. Humans are getting in on the action as well.
The reason for this is a confluence of factors that made books, and thus audiobooks, a tempting target for piracy, especially for those looking to make a quick profit.
Audiobook Piracy: Not Just for AI
According to an article by He Qigong at Sixth Tone, the TV drama The Protagonist has become a major success. Streamed over 1.18 billion times across its episodes, it has also brought significant attention to the original novel, written by Chen Yan.
But not all of that attention has been positive. Since the show’s release, local streamers have filled the hours of their broadcasts simply by reading the novel on stream. Though some streamers attempt to add some commentary or analysis as part of their reading, many just read the book itself.
To be clear, many of those streamers have had their accounts terminated. However, according to the article, this has been an ongoing issue in China. In April 2025, a novelist won 25,000 yuan ($3,700) in damages from a streamer who had read their novel on stream.
Fortunately, this does not appear to be as big of an issue outside of China, at least not yet. However, as one might expect, AI-generated audiobooks have become a growing problem, especially on YouTube. According to an article by Alexandra Alter at The New York Times, publishers have become increasingly frustrated with the lack of action from YouTube.
YouTube has a problem with low-quality AI-generated content broadly. It recently announced that it would be expanding its AI warnings and also its tool to prevent AI-generated likenesses of real people. However, when it comes to audiobooks, the company has been either unwilling or unable to take action.
Publishers are also in a difficult position. This is not a form of piracy that they are used to dealing with. They are, by and large, ill-equipped to deal with this kind of generative piracy.
So why is it becoming such an issue now? It’s the confluence of two factors.
Why Is It a Big Problem Now?
To understand why it’s a major deal now, we have to start in September 2017 with the launch of TikTok internationally. This marked a major shift in the way people consume video content. This created a gold rush for creating short-form vertical video content.
Facebook, Instagram, and YouTube would all release their own versions of the format, and it quickly became the most popular way to consume video content.
However, in November 2022, OpenAI released ChatGPT, kicking off the current explosion of consumer AI. This became a bigger deal in December 2024, when OpenAI released the now-defunct Sora, its video generation tool. Now it was trivial to create short-form video content with nothing but a text prompt.
This created a flood of AI-generated content in the space. While it’s unclear how many viewers were turned off, it made short-form video content much less profitable. That’s because, typically, any revenue from ads would be split between all the creators who had the required number of views. More creators meant less revenue, even if views were rising.
To be clear, this doesn’t change how heavily YouTube (and other platforms) push short-form content. It just means that, while the content gets more views, the creators earn less money.
As such, long-form videos became much more desirable. According to an article by Alexander Lee at Digiday, long-form video content generates well over 10x more revenue than views on short-form content. This is because you aren’t sharing the revenue with other creators, and you can place more ads in long-form content.
So, many are either pivoting to or pivoting back to long-form content. But that is a problem. Creating long-form content that people want to view is difficult. You could, for example, write an AI script and have another AI read it. Some certainly do that, especially in the educational space.
But books have an built-in audience. It’s easy to either read the book yourself on a livestream or have an AI do it. Even if you are completely unknown, try now that others will be seeking this content out. The audience will come to you.
While it’s a clearcut infringement, it’s also an infringement that YouTube is ill-equipped to handle. Though its Content ID system has been around for decades, it doesn’t do well when works are translated from one medium to another.
This was something that the Hbomberguy video highlighted. It’s easy to read from an article or a book and know that YouTube will not catch it. Content ID was designed to detect audio and video content that has been copied mostly directly, not audio based on someone else’s text.
When you combine this push for long-form content with the blind spot in various anti-piracy tools, you can see why creating or generating audiobooks has become an appealing option for many.
What Can Be Done?
There are no easy answers here. Publishers don’t have the relationship with YouTube that music labels and movie studios do. Historically, they haven’t needed it. YouTube is not a major platform for (legitimate) audiobooks, and publishers really don’t have a lot of leverage to push YouTube to develop the needed tools.
On the surface, this should be a relatively easy problem to solve. YouTube already generates captions for all the videos on its platform. Comparing those captions against a database of books should be more than doable. It would not stop all instances of piracy, but it would be a massive improvement.
But, as I said above, publishers don’t have the relationships with YouTube (and other live-streaming platforms). They didn’t really have a reason to build them. Now that’s an issue because they are caught flat-footed against a major piracy trend that threatens the burgeoning audiobook market.
While it would be nice for YouTube, Twitch, and other relevant platforms to step in and make a deal with this, it’s unlikely unless they are motivated to do so. This leaves publishers to develop their own solutions and, most likely, play Whac-A-Mole with pirated audiobooks.
Bottom Line
This case highlights how quickly the internet can shift and change. Just eight years ago, it would have seemed insane to imagine YouTubers uploading audiobooks or live-streamers reading them, especially en masse. The incentives were not there, and the technology wasn’t either.
Now, in the span of just a few years, it’s a major problem that’s only going to get worse.
In the end, I wouldn’t say that this is as serious of a moment as the early 2000s Napster era was for the music industry. But there is a strange echo as the publishing industry has been caught flat-footed against a major piracy trend, much like the music industry was then.
Addressing it will be difficult. It will involve creating new tools, forging new relationships, and, most likely, making changes to the industry itself.
If audiobooks are going to be a major part of the future of publishing, dealing with this problem as quickly as possible is going to be critical.
This is not something that publishers can or should ignore, no matter how tempting it may be.
Want to Reuse or Republish this Content?
If you want to feature this article in your site, classroom or elsewhere, just let us know! We usually grant permission within 24 hours.
