SoundStack, Audacy, and Sounds Profitable are teaming up for Clips: Podcasting's New Front Door on Wednesday, November 18 at 2 p.m. ET. Tom Webster will present new research from The Podcast Atlas on the Clip-Converted, podcast consumers who say clips send them back to the full episode and turn them into regulars of new shows. John Morris of SoundStack and Kate Hutchison-Rose of Audacy will also join the discussion. Registration is open now.
The last time I wrote about the AI data in The Podcast Landscape 2026, I asked you to guess which canceled show Americans named most often when we asked them about a favorite that got taken away before its story was finished. If you watched our recent webinar about this data, you’ve already seen the answer and the table that goes with it. For everyone else, spoiler: it was Firefly.
When I wrote the question over the summer, I actually called it the “Firefly question” because Firefly was the obvious show that got taken away from me too soon. It was so good that it even survived the later cancellation of Joss Whedon, the show’s creator. Of the 3,590 people who typed in a show, 81 wrote Firefly, more than twice as many as Friends (Friends ran ten seasons, so I suspect some of those folks are mourning a finale, which is a different survey). Fox made fourteen episodes of Firefly in 2002, aired eleven of them, and aired those out of order, and twenty-four years later it’s still the show Americans 18+ name most when you ask them what they lost.
Then we asked whether they’d watch an AI-made new season of that show, without the original cast or writers. Across everyone who’d lost a show, the answer was the coin flip I wrote about last time, 45% likely to watch against 41% unlikely. Firefly’s fans are not on the fence. Only 31% of them said they’d likely watch, and 61% said they were unlikely to, which puts the show our respondents named most among the least willing of the shows we can measure. Here are all of the shows named 20 or more times, ranked by the share of their fans who said they’d likely watch.
| Show named | Likely to watch an AI-made new season |
|---|---|
| The Walking Dead | 87% |
| Stranger Things | 68% |
| Game of Thrones | 62% |
| Friends | 58% |
| The Sopranos | 55% |
| My Name Is Earl | 54% |
| The OA | 49% |
| Mindhunter | 37% |
| The Big Bang Theory | 37% |
| Star Trek | 33% |
| Firefly | 31% |
| Blue Bloods | 27% |
| Yellowstone | 27% |
| Star | 15% |
| All who had a show canceled | 45% |
“Star” was written in on its own and most likely means the Fox drama. And mind the sample sizes: at this size a single figure can be off by 15 to 25 points, so please don’t quote The Walking Dead at 87% to anyone – it’s the general differences between the shows that is more salient to this conversation. Regular readers will recognize Mindhunter and The OA from my Tu-Dum article, where I complained about what happened to both of them, and now we know what their fans think of the machine option, which is “maybe” and “about half” respectively.
Age explains part of the spread. Stranger Things fans have a median age of 33 and Game of Thrones fans 34, and both groups are more accepting of AI hosts than the average person who lost a show (35% and 56%, against 21%). Blue Bloods and Yellowstone fans have median ages of 65 and 64, against 44 for the whole group, and fewer than three in ten of them would watch an AI version.
Grant Durando from Right Side Up, who co-hosted the webinar debut of this data with me, pointed out something on the webinar that the table doesn’t show, which is genre. Some of the shows near the top are fantasy or science fiction, Game of Thrones among them, and those shows already use a lot of computer-generated constructs anyway, because you can’t film Game of Thrones with a person in a dragon suit, which is why I sold mine. Some of the shows near the bottom, Blue Bloods and Yellowstone for example, are gritty realism (police procedural, ranch porn) which are not as reliant on a suspension of disbelief. The people who choose science fiction and the people who choose police procedurals are attitudinally different audiences, and I think that carries more water here than the age and gender differences do.
Firefly is the one that doesn’t fit either story. It’s a science fiction show, which by that logic should put it near the top. Its fans are 70% men with a median age of 40, and men aged 25 to 54 are the most willing demographic, at 57% to 58% likely to watch. By genre and demographics alone this should be an easy sell, and it’s one of the hardest.
I mentioned last time that how people already feel about AI hosts predicts this answer better than any demographic we measured, and that’s what is doing the work here. Only 15% of Firefly’s fans call AI hosts acceptable. I’d treat that 15% gently, since it rests on a base of 81, but it’s the most plausible explanation I can find for a fan base that runs against its own profile.
My own theory, and it’s only that, is that the Firefly result is probably idiosyncratic to Firefly. We specified “without the original cast or writers,” and Firefly’s original cast is still around and still vocal about getting back together, which even a casual fan would have read about. They are also currently collaborating on a new animated version of the show that will feature their original voices. Every show’s fans lost their cast and writers in that question, but a Firefly fan may well have been answering with the real reunion in mind. Firefly is also a show whose fans quote the dialogue back to each other at conventions, and a show with only fourteen episodes has very little in it that isn’t those people.
Monthly podcast consumers are more open to an AI season than people who’ve never consumed a podcast (52% likely against 32%), but I wouldn’t read that as a preview of where everyone is headed, because I’d guess people who consume a lot of things are early adopters of many things.
The point I’d take from all of this is that the podcast industry needs to do more content research. There is a lot of advertising and sales research in this business, and lord knows I’ve contributed to it in my own way. The best way to know how your audience is going to react to AI augmentation, AI assistance, AI generation, or any kind of AI involvement at all in your podcast, is to ask them. The Firefly result is what bears that out. You might have looked at the character of the other shows on that list, the genres and the ages of the people who named them, and guessed that Firefly’s fans would be fine with an AI-made season. You’d have been wrong, because that audience is their own dog.
I know from a long career in doing it that content research is the key to building a strategic moat around your shows and your IP, because nobody can tell you what your audience will accept except your audience. I do hope we see the day when podcasting does a lot more of it.
The whole AI and Podcast Consumers report, which covers a good deal more than television, is free right here. Forward it to whoever you’ve been arguing with about Firefly.
Fox took the sky from me in 2002, and I’m with the 61% who are unlikely to watch an algorithm try to give it back. Shiny.
Thank you to Right Side Up, sponsor of AI and Podcast Consumers, and to the wonderful sponsors of The Podcast Landscape 2026: American Public Media, BetterHelp, NPR, ESPN Podcasts, SiriusXM, and DAX, as well as our research partners at Signal Hill Insights. The Podcast Landscape 2026 comprises a sample of 5,473 US adults, 18 years of age and older, weighted to the US Census to represent the US population.
