by Tom Webster
A tease of upcoming audience research, including whether folks would watch an AI revival of a cancelled favorite show.

Well? Would you? Go with your gut reaction, and I'll wait.

In our latest research project, The Podcast Landscape 2026, we asked 5,473 Americans a bunch of questions about AI and podcasts, and a question that superficially has nothing to do with podcasts: "Have you ever had a favorite TV show get canceled before you felt its story was finished?" Sixty-five percent said yes, so if you've been holding a grudge against a network executive, you are not alone.

We then asked those people a second question: "Imagine AI could create a brand-new season of that show — new episodes in the original's style, without the original cast or writers. How likely would you be to watch it?"

I built the questions that way on purpose. If you ask people whether they'd watch an AI-generated show, or listen to an AI-generated podcast, most of them will say no, because they think they should say no, and that tells you who they'd like to be rather than what they'd do. Hypotheticals are my least favorite thing to ask in a survey, and I've seen other data on whether people would watch AI-generated content that I think is useless for exactly that reason. If you had asked people whether they'd get into a car with a stranger before Uber, you'd have gotten a 95% no. And the 5% are interesting people.

So I anchored the question in something specific and a little painful: loss. If you put people into the frame of reference of a show they loved, one that was taken away before its story was finished, would that change their default reaction? The hypothesis, of course, was that a real loss would either shake people out of the aspirational answer and into something closer to the truth, or make them double down into a HELL NO THE ROBOTS AREN'T TOUCHING MY SHOW.

I'm going to give you the overall answer here, with the caveat that it varied wildly by the actual shows named. Among everyone who'd lost a show, 45% said they would be at least somewhat likely to watch an AI-generated season, while 41% said they were at least somewhat unlikely, with 14% in the middle, which is about as even a split as survey research ever hands you. Truly a coin flip, with not many people on the fence. It's worth pointing out, though, that the passion in the data trends towards rejection, since 29% said very unlikely compared to 19% who said very likely.

What predicted the answer better than any demographic we measured was how people already feel about AI hosts. Among those who call AI hosts acceptable, 80% would watch an AI-made season, and among those who call them unacceptable, 28% would. A question that sounds like it's about television, in other words, is also a question about how people feel about the machine.

Let's not read the 45% who were at least somewhat likely as a green light, of course. We asked whether people would watch, not whether they'd like it, and "would watch" is a low bar that can't separate the fan who'd be thrilled from the one who'd tune in once to see how bad it is. I am sure I could ask a question about the likelihood of watching a show about Green Lantern, but there is a world of difference between the recent HBO series Lanterns, and whatever that monstrosity Ryan Reynolds made while he was fishing around for which superhero he was going to be. I am pretty sure his Green Lantern movie was written by Claude Sonnet 1.1 with CGI provided by whatever made those early Will-Smith-eating-spaghetti videos.

Would I watch an AI-generated version of a favorite show? My own answer is nuanced, which I realize is the least useful thing a researcher can say about a survey question: it wouldn't bother me much if elements of a revival were AI-generated, but the humans better be writing and voicing the show. Luckily, we asked a few more questions to get a bit more granular about this topic.

For one thing (since I don't love hypotheticals), we asked everyone who'd lost a show to type in its name, and 3,590 of them wrote in 1,888 different shows (!), which anchored the follow-up question and allows us to see which shows people named, and how the answer to that second question differed from one show's fans to the next. We also asked a bunch of questions about using AI to discover podcasts, and we are going to go through all of that data with Grant Durando from our friends at Right Side Up in a free webinar on Nov 9 on AI and podcast consumers. You are going to want to register for that one. Bring your AI notetaker, I am sure it will amuse them.

In the meantime, a bonus question: which show do you think was named most often? For what it's worth, I had a show so clearly in mind when I wrote the questionnaire that I named this question after it, and it turns out I was right. Make your guess now, and bring it to the webinar along with your gut reaction and your grudge.

Thank you to Right Side Up, sponsor of AI and Podcast Consumers, and to the 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.