“Synthetic Stories” — an AI-generated podcast launches and makes us think of AI audiobooks.

A new podcast, Synthetic Stories is new in two ways. It just launched, for one, and it is the first podcast (as far as we know) produced entirely by Artificial Intelligence (AI). Created by production studio This Is Distorted, which calls the show a “truly groundbreaking podcast series,” every aspect of the show is synthetic — the script, the sound design, the music composition, the artwork, the voice performance, and the promotional materials accompanying the release. (See the description and listen HERE.)

The producers promise “a truly unique and engaging listening experience.” We agree that it is (as far as we know, for now) unique.

To our ears the experiment is a clear success. At the same time, when you know it’s a synthetic production, you can hear the ways in which it is. Coming into it knowing nothing, one might think it’s a competent podcast with exceptionally steady narration, and a music soundtrack that remains simple and mostly unchanging throughout.

If the music is a giveaway, the hosting experience is more convincing. But (again, to our subjective ears), the vocal track sounds less like the inflections and conversational style of a podcast, and more like the narrative style of an audiobook.

With that in mind we visited Apple’s page on digital audiobook narration technology. The company is far along on making synthetic audiobook narration sound good. (Listen HERE.) Steady cadence and relatively flat pitch is a combination that works well for audiobooks, even if star performers in the genre are sometimes more modulated and dramatic.

In both cases — articially produced podcasts and audiobooks — the advancing technology poses a shadowy threat to human performers. And a potential cost-efficiency advantage to producers.


Brad Hill