Amazon’s Audible just launched a new feature where listeners have real-time conversations with AI-powered literary characters. The experience begins with the character setting the scene, assigning the listener a role and establishing stakes. The listener responds in their own words, and then the character reacts.

Audible calls it “Interactive Story.” It is the first time a major audio platform has invited listeners to contribute to a story experience.

The feature lands squarely in the middle of the most contested legal questions about AI and copyright law. Could Audible expand the new feature to its entire catalog without infringing copyrights? And when listeners participate, do they create their own copyrights?

The answers are unclear, and the law is still catching up.

Where AI And Copyright Law Stands

U.S. copyright lawsuits against AI companies reached 140 cases in September, up sharply from 45 cases just over a year ago. They fall into two categories.

The first asks whether AI companies infringe copyrights by using protected works to build the model. The answer depends on the particular facts of each case, so courts have reached different outcomes.

A California federal judge ruled that Anthropic’s use of lawfully acquired books to train its AI qualified as fair use and not infringement. But the judge ruled that Anthropic’s use of books downloaded for free from pirate sites was infringement. Anthropic later settled the case for $1.5 billion, Reuters reported .

A different California federal judge ruled that Meta’s use of books it downloaded for free from “shadow libraries” to train its AI model qualified as fair use and was not infringement.

But a Delaware federal judge ruled that Ross Intelligence’s use of Thomson Reuters’ editorial content and annotations to train its AI legal-research tool was infringement.

The second category is whether AI outputs infringe the copyrighted works the AI learned from. Courts have not answered this question.

For example, Disney and Universal sued Midjourney in California federal court, arguing that Midjourney’s AI image-generating service “functions as a virtual vending machine, generating endless unauthorized copies of Disney’s and Universal’s copyrighted works.” Warner Bros. Discovery later joined Disney and Universal in the lawsuit.

Midjourney says its model does not infringe any artwork. “When a user enters a prompt, the model doesn’t retrieve an image from storage; rather, a new image is synthesized from scratch, guided by user instruction and the statistical patterns encoded within its artificial neurons,” Midjourney said in its answer denying the claims. The case is pending and the judge has not ruled on any claim.

And Getty Images sued Stability AI in California federal court, claiming in part that Stability AI copied more than 12 million photos without permission to train its AI model and that the resulting output infringes Getty Images’ copyrights. Stability AI denies the copyright infringement claim. The case is pending and the judge has not ruled on the issue.

Audible’s Interactive Story feature touches both categories. The AI generating a character’s responses was trained on something. And what it generates in real time is output. It isn’t clear yet whether either could constitute copyright infringement if Audible were to expand the feature beyond Audible Originals.

What Audible Is Launching

Audible says creators define the foundations of each Interactive Story, like opening scenes, turning points and character persona. Audible’s product design will keep the story faithful to the creator’s vision throughout.

But Audible is not launching the feature for its entire catalog. At least not right now.

Audible is starting with its own production of Bram Stoker’s Dracula . The Interactive Story feature follows the character Renfield, Dracula’s unpredictable, tormented servant.

Then on November 5, Audible will use the feature for Exoplanet , an Audible Original written by author Benjamin Percy.

Audible controls the rights to its production of Dracula , a public domain work, and Exoplanet is an Audible Original . But Audible’s catalog contains hundreds of thousands of titles owned by others.

For example, the audiobook of Stephen King’s It , published by Simon & Schuster Audio, is on Audible. So is the audiobook of Gillian Flynn’s Gone Girl , published by Random House Audio. Audible does not own those works. Instead, Audible has agreements (licenses) that give it the right to sell and stream them.

Those licenses almost certainly do not contemplate an AI system generating new dialogue for King’s or Flynn’s characters in real time.

Whether that use falls within an existing license or requires a new license are questions existing agreements probably do not answer. By launching with its own works first, Audible is testing demand before navigating those conversations and legal hurdles.

The User Content Question

Other questions include whether the listeners create copyrights by using the new feature, and if so, who has the right to use them. No court has answered these questions.

Under U.S. copyright law , the person who creates an original work owns it automatically. No registration, filing or deal is required.

When a listener responds to Bram Stoker’s Dracula character Renfield in their own words, crafting specific dialogue and making narrative choices, they are arguably creating original expressions. The default rule says the listener owns that original expression.

But the listener’s contribution doesn’t exist in a vacuum. It’s woven into a story framework Audible created, responding to Audible’s characters and through a system Audible built.

Under U.S. copyright law, a new work that builds on existing copyrighted material like a sequel or adaptation is called a “ derivative work .” The creator of a derivative work owns only the new material they contribute. That suggests a listener could own their own contributions, while Audible keeps its story, scripts and other copyright ownership.

In practice, Audible’s terms of use will likely determine who can use what listeners create. Audible incorporates into its terms the Amazon.com conditions of use , which state that users “grant Amazon a nonexclusive, royalty-free, perpetual, irrevocable, and fully sublicensable right to use, reproduce, modify, adapt, publish, perform, translate, create derivative works from, distribute, and display” user content and material “throughout the world in any media.”

If that provision covers the Interactive Story feature, then even if listeners own copyright in their contributions, Audible can use them in virtually any way.

Amazon and Audible did not respond to a request for comment.

The AI and copyright litigation landscape is moving fast—140 U.S. lawsuits and counting, a $1.5 billion settlement and courts still working out the foundational rules.

Audible’s Interactive Story feature adds a new layer about what AI and humans create together in real time and who owns the result. The answers will matter well beyond audiobooks.