Don’t Trust The Best Of… Lists For Podcasting

The end of the year brings the inevitable BEST OF… lists for all forms of media. Best Books, Best Movies, Best TV Shows, and, of course, Best Podcasts.

“Best of” lists are criticized for being subjective, biased toward popular or mainstream items, and reducing complex art or products to simplistic, arbitrary rankings. They often omit worthy, less-known options, prioritize clicks over quality, and create “winner-take-all” environments that overlook the nuance of individual taste.

According to an NPR article, Best Of…Lists are essentially personal opinions disguised as objective rankings. They reflect the tastes of a few critics or algorithms rather than a true universal “best.”

More egregiously, high-ranking items are often just popular, familiar, or aggressively marketed, rather than the best quality.

As Linda Holmes of NPR once pointed out: “The entire ‘How can this be on your list and not that?’ sequence that repeats itself endlessly over and over at this time of year is based on the assumption that if my list would differ from your list, then I naturally conclude that your list is flawed in a way that requires an explanation.”

Another inbred flaw of these lists is that they often focus on mainstream content, ignoring overlooked masterpieces or smaller, independent artists. In podcasting, for example, so many writers or social media gadflies publish the Best Of List as a recycled version of the Top Ten podcasts ranked by downloads. Song Exploder is the default setting for the best music podcast, while One Song, which dissects one iconic song every episode, and provides a deep, humorous, and technical analysis of a single track’s production, history, and cultural impact goes largely unnoticed.

Then there’s the “Winner-Take-All” Mentality that forces a competitive framework onto art, as if rating podcasts is like using baseball statistics to evaluate a player. These lists create an environment where ranking 10th seems worse than 1st, despite both being high quality. Even A “Top 10” is often arbitrary—there might be 20, or only 3, truly “best” items in a given year.

After all, there are 40,000 active podcasts, culling a top ten from that massive list is essentially needle meet haystack. Movie production has slowed to the point that a Top Ten list of the Best Movies of the Year actually includes about ten percent of all the movies made that year.

Finally, there’s the eventual clickbait aspect to Best Of…Lists. Critics note that these lists can be a form of lazy journalism, designed to generate web traffic (clickbait) rather than curate thoughtful recommendations. Indeed, Best Of…Lists are almost a guaranteed winner for any digital publication or the few print publications still in existence.

Here’s writer Amy Bernstein on Best Of…Lists in a 2023 article in Medium. “I’ll never understand the best-of mania that grips America at the end of the year. As if the preceding 11 months were nothing more than a marathon we’ve all been running, and only a handful of us can claim a spot at the finish line.”

What does BEST mean in an AI world?

That focus on recommendations prompted a thought experiment for me. When someone queries an AI model — ChatGPT, Copilot, Gemini — for podcast recommendations, does the result by the AI model include any independent podcasts?

Before answering that question, I took a remedial lesson in artificial intelligence (AI). I learned that AI answers recommendation questions by using Machine Learning to analyze vast user data and patterns (like viewing/buying habits) to predict what you’ll like, employing techniques like collaborative filtering (you’re like others) or content-based filtering (you like X, so here’s more X), synthesizing information from trusted sources, and offering structured, context-aware suggestions for products, movies, or content, though its suggestions rely on data and lack true human judgment.

AI chatbots like ChatGPT, Perplexity, or Copilot are not just conversational tools – they’re intelligent search interfaces. Unlike traditional search engines that return a ranked list of links, AI chatbots generate direct answers by synthesizing information from various sources and matching it to the context of the query.

Their key advantage is understanding user intent – chatbots don’t look for keyword matches but instead try to respond precisely to the query, considering the conversation context, question style, and even previous interactions. This allows them to not only provide information but also recommend specific products, services, or solutions – just like a consultant or salesperson would.

I know that was tough. Here’s the result I received when I asked AI to provide me with the Best Podcasts of 2025.

Top Ranked & Most Popular (2025)

These shows consistently lead national charts by listenership and cultural impact. Therefore, the AI associates downloads and listens with “The Best.”

Notice what happens when I revise my AI query and ask it to name the BEST INDEPENDENT PODCASTS OF 2025.

Storytelling & Human Interest

• Rumble Strip: Produced by Erica Heilman from her home in Vermont, this show is widely considered a masterpiece of the medium. It features deeply intimate interviews that capture the “idiosyncrasies of the human condition”.
• The Life Shift: Named Best Overall Podcast of the Year at the Ear Worthy Awards, this show features candid conversations about pivotal, life-altering moments.
• Multispective: Recipient of the Best Life Lessons Podcast award, focusing on how unique personal experiences shape everyday life.

Specialized & Niche

• Ear Hustle: Co-produced by Nigel Poor and Earlonne Woods, this show provides a raw and often funny look at life inside prison.
• The Indie Hackers Podcast: Essential listening for entrepreneurs and developers looking to build profitable businesses without venture capital.
• The Sound Session: A 2025 favorite that explores the impact of sound on the world from both artistic and scientific perspectives
• Verbal Diorama: Awarded Best Movie Podcast in 2024-2025 for its deep dives into the history and legacy of films.
• Imaginary Worlds: A sound-rich exploration of science fiction, fantasy, and speculative fiction.

As you can plainly see, the AI returns a reasonable and altogether accurate result based on my query. However, why is it that a query to an AI that does not specifically include the term “independent” does not return any independent podcasts when asked for BEST PODCASTS OF 2025?

What’s Best for the consumer when they request The Best Of…List?

Does a podcast listener care if the best podcasts are network-produced or independent? In this case, no. However, when the AI excludes independent podcasts from the AI query result, that’s a problem for podcast listeners and podcasters.

It’s a problem for podcast listeners because they’re not receiving an accurate answer. What if a person asks an AI to name the greatest baseball players of all time and the AI only returned results on white baseball players, excluding Black, Asian, and Latino players?

Without including independent podcasts in the query result, podcast listeners are not receiving accurate and complete information by which to make a decision on our next podcast options. In fact, the AI has already decided that “what is most popular is the best.”

“When these AI-driven Best Of…Lists simply return the top ten podcasts in downloads,” notes data analyst Lawrence Carruthers in an interview, “You can skip that step of discovering the Best Of Lists. The AI chatbot doesn’t have a method for understanding a user’s individual artistic criteria, so it defaults to most popular.”

The AI result on BEST PODCASTS is an issue for independent podcasters because they are excluded from a valuable and much-used podcast discoverability tool. After all, asking “the computer” for recommendations is becoming more prevalent every day.

Can anything be done about this inequity? Do we not recognize that technology has already captured our attention, viewpoints, political beliefs, and now our media choices?

Why can’t an AI query on BEST PODCASTS return a result that includes INDEPENDENT PODCASTS, even as a distinct category?

I apologize if this is one of those frustrating stories for readers when the writer lays out the problem and has few solutions. I’m not a technology expert. I just discovered how to increase my font size on my iPhone, so I’m not I.T. Help Desk material.

Podcast consultant George Witt says: “My outrage is at DEFCON 5. This is just another example of how independent podcasts are excluded from the podcast industry adult’s table and are relegated to the kid’s table. You know, visible by not audible.”

Recently, Spotify, expanded its Prompted Playlist beyond music to now include podcasts. Launched earlier this year in beta, this feature lets you steer the algorithm by creating playlists based on your listening history and real-time signals on Spotify, including trends and charts, so the playlist reflects both your tastes and what’s happening in the world around you.

Rolling out in English to Premium users in the U.S., Canada, U.K., Ireland, Australia, New Zealand, and Sweden, Prompted Playlist (beta) can now help listeners discover new shows, go deep on a specific topic, or build a curated lineup to match their mood.

“Podcast fans are always looking for their next great listen,” said Lizzy Hale, Spotify’s Global Head of Podcast Editorial. “Prompted Playlist makes discovery feel effortless and personal. For creators, it unlocks powerful new opportunities for discovery, bringing both back catalog and new episodes to audiences who are actively signaling what they want to hear.”

As data analyst Lawrence Carruthers emphasizes: “Most podcast hosting companies and podcast apps have the ability to curate podcasts based on your personal preferences and tastes. A Best Of…List is simply a reflection of someone’s tastes or what they think the readers of their publication of social media space would like. With AI, these top ten lists are now outsourced to a non-human. For some, AI offers more objectivity, yet an AI can only deliver a list based on the parameters of its programming.”