The oldest digital library in the world, Project Gutenberg, has transformed thousands of ebooks into audiobooks using AI—bypassing the longer (and more expensive) process of hiring a human reader to do the job. It’s exactly the kind of AI application that actors, who are currently on strike in the US for the first time in four decades, fear may endanger their careers.
Using synthetic speech technology, Project Gutenberg, in partnership with Microsoft and MIT, transformed more than 5,000 of the ebooks in its free and open access collection into audiobooks. Available on Spotify, Apple Podcasts, and Google Podcasts, the library now contains an eclectic mix of titles such as The War Poems by Siegfried Sassoon and England, My England by DH Lawrence, all of which are open-license.
The collection is read in a male voice—youthful, with an American accent—in a delivery that (to this writer’s ears) is fairly naturalistic. Aside from an occasional stilted pause, it would be difficult to recognize it as AI-generated speech.
Project Gutenberg’s AI audiobook reader is no match from Stephen Fry
What the reading does lack is the versatility and depth that human actors bring to the same work. It’s a far cry from, say, Stephen Fry, who created a unique voice for dozens of characters while reading the Harry Potter audiobooks. But Project Gutenberg’s audiobooks are free, the process is quick, and the text-to-speech system used is scalable. The threat that such technologies pose to acting work is clear.
Project Gutenberg’s methodology was published in a report (pdf) titled “Large-Scale Automatic Audiobook Creation,” on Sep. 7, and the code for the project is open source. In the introduction, the research team mentions that “audiobooks can take hundreds of hours of human effort to create, edit, and publish.” Later on, the authors say: “Our system allows users to customize an audiobook’s speaking speed and style, emotional intonation, and can even match a desired voice using a small amount of sample audio.”
Project Gutenberg is a volunteer-run, non-profit organization whose mission is to “encourage the creation and distribution of ebooks.” Ironically, the archive itself has been used to train AI models like Meta’s Llama.
Quartz has contacted Project Gutenberg for comment.
Actors on strike want to protect their livelihoods from AI
AI has been a central bargaining point in the current US actors’ strike. The Screen Actors Guild and American Federation of Television and Radio Artists (SAG-AFTRA) has been on strike since July 14, demanding not only better wages and shares of residual, but also protections from AI technology that can create a “digital replica” of an actor without their consent. Digital replicas include a person’s image and as well as their voice.
On Sep. 5, Fran Drescher, the president of SAG-AFTRA, wrote in a letter to the union’s members that protections from AI were not only “righteous and fair” but also necessary to guard against its “existential threat” to their work, especially given the lack of regulation.
“A great deal of our members’ work in this space is voiceover, and the capacity to cheaply and easily create convincing digital replicas of performer voices is already here and widely available,” Drescher wrote in the letter. “Without protections, not only will this be the future of how voices are recorded for video game characters, but your own voice recordings will be used to train the AI systems that replace you.”
The fact that the Project Gutenberg method can use a single audio sample to build out an entire text-to-speech voice is case in point.
Most voice actors aren’t in a union, though. Only about 18% of those in the trade are SAG-AFTRA members, according to data from a 2022 survey conducted by the National Association of Voice Actors. More than 64% in the survey said that they were non-union. Given the industry’s relatively less organized nature, it will only get easier and easier for AI to take over tasks that once only humans could do. Uniting their voices might be what it takes to ensure that voice actors can keep their rights to using them.
📖 Quartz Obsession | Audiobooks: Reading allowed
🎥 Movie directors got an AI deal with studios—but striking writers still have no such promises
✨ The Marvel universe has new avengers: Visual effects workers want to unionize