AI could enable a golden age of audio piracy

Everyone knows the story: In the early 2000s, peer-to-peer (P2P) file-sharing platforms like Napster, Kazaa, and LimeWire helped digital music piracy proliferate, ultimately decimating recorded music revenue for years. Almost a decade later, a similar, though not quite as dramatic, fate befell book publishing. 

Fortunately for authors, it was much harder to illegally copy and distribute books, which made it possible for digitisation and monetisation to happen concurrently — even if it was at the expense of higher margins. In the music industry, it took years for the law to catch up to the innovation, which meant digitisation far outpaced monetisation. 

We find ourselves at a similar inflection point today as technology — particularly AI — rapidly outpaces legislation, bringing digital audio again to a point where innovation happens at the expense of creator compensation.

From fake artists to fake books

For years, journalists have covered music streaming fraud, but there have been virtually no repercussions for fraudsters or active efforts to address the issue until recently. Besides Deezer and Spotify’s recent self-described efforts to curb fraud through changes to their royalty payout structures, on March 21, a Danish man became the first person convicted of streaming fraud, having both infringed other artists’ copyrights and used bot accounts to drive up streams on the tracks he uploaded. Notably, much of this fraud happened between 2013 and 2019, well before the AI boom of the post-pandemic. That raises the question: How much more sophisticated has streaming fraud gotten with the assistance of AI?

Book publishing may already be giving some clues, as some authors are reporting that AI-generated books published under their names or oddly similar names are cropping up on popular sites like Amazon and Goodreads. Considering all you really need to submit an audiobook is to have a print or digital version available for sale, it is only a matter of time before AI-generated text-based books become audiobooks — especially as AI celebrity voice generators get more and more realistic. Without legislation that can both address current issues and also account for issues that may arise in the future, Amazon and Goodreads’ AI moderation game will long be a tireless whack-a-mole effort. Rather than resisting it, platforms may want to consider creating secondary markets for AI-generated content, working with authors to make disclaimers clear and ensure that they are appropriately compensated for any derivative works that may be generated.

Digital audio bootleg platforms rise again

Now that the digital monetisation seal has been broken, it is not likely we will return to the days of P2P file-sharing. Whether through subscriptions, advertising, or data capture, nothing on the internet is truly free anymore. However, the risk to the creator’s bottom line — and to the platforms expending resources on content removal while battling copyright infringement — remains. The top concern is no longer consumers pirating audio content, but fraudsters duping consumers into consuming “fake” content, or redirecting revenue from real creators to themselves.

The shift to streaming subscriptions proved that sometimes legal intervention works to curb piracy and encourage innovation. At the end of the day, most users will go to the platforms that have the most content at the lowest price, and AI offers the ability to scale content at an unprecedented rate. For now, considering it is much easier to replicate an author’s voice with AI than it is to create a hit song with AI, audiobooks seem much more vulnerable to audio piracy this time around.