by Gavin Gaddis

I’m typing this from behind a locked door. We miscalculated and ran out of seed for the bird feeder. There’s a Hitchcock movie brewing on my patio. I need to grab my keys and make a run for the car, maybe there’s time to buy more seed before they descend? Here’s the week’s news before I bolt.

Tom Webster returns to the AI data in The Podcast Landscape 2026, a study of 5,473 U.S. adults. The survey asked respondents to name a favorite show that was canceled before its story was finished. Of the 3,590 people who typed in a show, 81 wrote Firefly. That is more than twice as many as Friends. Respondents were then asked whether they would watch an AI-made new season without the original cast or writers. Across everyone who lost a show, 45% are likely to watch and 41% are unlikely. Only 31% of Firefly fans are likely to watch, and 61% are unlikely.

Webster tests a few explanations for the spread between shows. Age accounts for part of it. Fans of Stranger Things and Game of Thrones have median ages of 33 and 34, and both groups are more accepting of AI hosts than the average person who lost a show. Fans of Blue Bloods and Yellowstone have median ages of 65 and 64, and fewer than three in ten would watch an AI version. Grant Durando of Right Side Up raised genre as another factor on the webinar debut of the data. Fantasy and science fiction already lean on computer-generated effects, while police procedurals and ranch dramas do not. Firefly fits neither story. It is science fiction, and its fans are 70% men with a median age of 40. Men aged 25 to 54 are the most willing demographic overall. Only 15% of Firefly fans call AI hosts acceptable, though Webster cautions that figure rests on a base of 81.

Webster argues the lesson for podcasting is to do more content research. Monthly podcast consumers are more open to an AI season than people who have never consumed a podcast, at 52% likely against 32%. He does not read that as a preview of where everyone is headed. “Nobody can tell you what your audience will accept except your audience.” The full AI and Podcast Consumers report is free.

Analyst Adam Bowie examines RAJAR’s Autumn 2026 MIDAS survey. It is a recontact study fielded over two weeks in September with 1,717 former RAJAR participants aged 15 and older. The survey finds 96% of U.K. adults listen to some form of audio each week, averaging 27.8 hours. Weekly podcast reach sits at 26%, where it has been for about a year. Bowie notes that RAJAR’s regular quarterly survey uses a different method and puts weekly podcast listening at 30.7%. Just 11% of podcast listening happens in the car, compared with 25% of radio listening. Word of mouth remains the top way people discover podcasts, followed by social media. Most listeners hear three or fewer episodes a week. Only 6% hear 11 or more.

Bowie also covers separate RAJAR research drawn from its monthly Podcast Tracker. Director of Research and Audiences John Carroll presented it at the ASI audio conference in Venice. It finds 34% of podcast consumers move between audio and video in the course of a week. Another 37% listen to audio and 29% watch video. Among video podcast consumers, 20% say they mostly watch. Another 35% mainly listen and only occasionally look at the screen, and 43% split their time fairly evenly. Household chores are the top activity while consuming video podcasts at 41%. Doing other things on a phone follows at 37%. Bowie compares the behavior to daytime television that keeps people company while they do something else.

RAJAR’s own write-up of the research adds that podcasts deliver 105 million hours of listening a week in the U.K. Around half of respondents, 49%, say seeing hosts’ reactions and body language is what video brings to a podcast. Most podcast users, 62%, say they would still consume the same content if it were available only as audio. “In practice, much video podcasting is better understood as audio with a visual layer, rather than television by another name,” RAJAR writes. The organization notes that younger audiences are the pressure point, with video much closer to the default for ages 16 to 24.

Triton Digital President and CEO John Rosso responds to a Podnews investigation into copied shows on Spreaker. Podnews editor James Cridland reported in October that a free Spreaker account could import any RSS feed, keep the copy updated hourly and turn on monetization. Podnews found 1,778 shows in the Podcast Index that appeared to be copied onto Spreaker and listed in Apple Podcasts. Rosso says duplicate feeds are a small fraction of the podcasts on the platform and many are authorized. He acknowledges that a limited number are unauthorized copies.

Spreaker will now email the owner listed on the original feed whenever a show is imported. Before publication, it will check the Apple and Podcast Index APIs to see whether a feed already exists elsewhere. Any feed flagged as a possible duplicate will need an extra approval step, including a check for a valid redirect, before it can be monetized. Sign-up and monetization emails will carry added reminders of the platform’s rules. “We’re targeting to have these changes in place by year’s end,” Rosso writes. Podnews has also drafted a Code of Practice for podcast migrations. Blubrry, Captivate, RedCircle, RSS.com, Spreaker and Transistor have committed to follow it.

A second Podnews investigation this week looks at a different platform. Cridland reports that Muse, the AI assistant Meta announced in September, has podcast production and hosting built in. It serves both audio files and RSS feeds from muse.ai/podcasts. The reporting began with a pitch from a “Charles Mayfield” of Professional Development Inc. about a network called Education Across the States. Its feed held 1,503 episodes published between October 19 and November 7. That totals more than 441 hours, or about 23 hours of audio per day.

In written replies, Mayfield said the network has no deal with Meta, gets very little analytics from Muse and has no monetization features through it. Cridland found no evidence that Mayfield or his company exist. The supplied press photo carries a watermark identifying it as AI-generated. Dublin City University told Podnews it will not offer students a placement role with the network that it had provisionally accepted. Podnews counts 22 podcasts hosted on Muse so far, several presented by AI hosts named “Alex” and “Jordan.”

Your deals or ours. Most audio buys lose control the moment your audio file goes out. Podscribe SmartServe takes it back.

Two ways in:

Apply AI Optimize to both. Let your conversion data help serve the impressions most likely to drive results. Randomized holdouts then show which conversions your ads actually caused.

While they may not be top story material, the articles below from this week are definitely worth your time: