Episode 88 (AI and the Human Voice)

The May 2025 episode is the first to tackle the red-hot topic of artificial intelligence, or AI. Paul’s guests are speech-to-text specialist Michael McAuliffe and dialect coach Tanera Marshall. Paul, Michael, and Tanera, along with podcast co-producer Cameron Meier, discuss AI in the context of human speech in general but especially acting and singing.

Michael, a linguist, brings AI knowledge to the conversation, while Tanera, by virtue of her being the dialect coach for The Brutalist, lends a unique perspective. It was The Brutalist that generated controversy by using AI to tweak the Hungarian-language dialogue of its actors. For more perspective on the film, see Cameron’s review and the videos toward the bottom of this page.

Among the issues discussed are the definitions of AI, the technological achievements and current limitations of computer “intelligence” to alter the human voice, the future of dialect coaching and acting itself, the legal and financial issues involved with using actors’ voices to train AI models, the moral and practical differences between AI and older cinematic technological illusions (editing, looping, rear projection, green screen, CGI, motion capture, de-aging, etc.), and the use of AI in music. To demonstrate the latter, the episode includes a portion of a song that was recently created in AI. To hear that entire song, visit Suno. (Thanks to Filmapalooza, Duane Trammell, and Jahil Nelson for use of the song.)

Michael McAuliffe is an adjunct professor in the Linguistics Department at McGill University, focused on creating and improving computational models to enable phonetic analysis of speech. He is the maintainer for the Montreal Forced Aligner, a tool that segments audio files into speech sounds based on a text transcript. His other research interests lie in how listeners adapt to novel speakers and accents, and phonetic analysis of dialects of English as part of the SPeech Across Dialects of English (SPADE) project. Also see the Montreal Forced Aligner.

In addition to her previously mentioned achievements, Tanera Marshall is a senior editor of the International Dialects of English Archive, or IDEA. She is a voice and accent specialist in theatre, TV, and film, and an associate professor at the University of Illinois at Chicago’s School of Theatre and Music. Her current and upcoming productions include The Brutalist, Train Dreams, Oh.What.Fun., Ann Lee, The Morning Show (season four), The Bear, and Chicago Fire. Recent credits include Hunger Games (2024), Asteroid City, Widows, Queen & Slim, Dungeons & Dragons, Fargo (season four), Station Eleven, Army of the Dead, Public Enemies, The Immigrant, and the Divergent series. Stage credits include Broadway tours of Hamilton, Billy Elliot, and Dirty Dancing, as well as productions at Steppenwolf, Court, and Timeline theatres in Chicago. For more information on Tanera, go here. And don’t forget to visit her page on IDEA.

For a demonstration of the real-time accent modification mentioned in this episode, visit Krisp. Also check out Natural Reader, a company that uses AI to turn text into speech for films and audiobooks. Notta and ElevenLabs provide similar services.

Related content:

For more perspective on Adrien Brody’s performance in The Brutalist, check out this interview on TikTok.

Finally, see what our old friends Gillyanne Kayes and Jeremy Fisher have to say about AI:

(Bach’s Cello Suite #1 in G Major BMV 1007 Prelude (by Ivan Dolgunov) is courtesy of Jamendo Licensing.)

The post Episode 88 (AI and the Human Voice) first appeared on Paul Meier Dialect Services.