In February 2026, we asked seven AI tools about emerging voices in the podcast industry and which podcast leaders they cited most in their results. That article gave us a useful, if imperfect, snapshot of how AI “sees” the podcasting space. But a lot can happen in a few months in AI tool development, and we thought it was worth running a version of this exercise again, but by asking a more timely question: what are the best podcasts of 2026 so far? Since we’re now halfway through the year when podcast critics publish these “Best Podcasts of the Year (so far)” pieces and listicles, this seems like a good time to check in on AI platforms and see what surfaces.
It’s important to note that while the question we’re asking is pretty straightforward, the responses from these AI tools are not a definitive ranking of the best podcasts in the industry right now. What this is, instead, is a picture of which new podcasts have built enough of an online presence to get cited when an AI tool is asked this particular question. If you want to know how to make your podcast more discoverable in search engines and AI tools, then be sure to check out our podcast growth services or subscribe to our free newsletter for more insights and information like this.
How did we run this exercise?
On June 2, 2026, we started with this question: what are the best podcasts of 2026 so far according to AI tools?
We built on this and created a prompt to query the following AI tools (with their LLMs in parentheses): ChatGPT from OpenAI (GPT-5.5), Claude from Anthropic (Sonnet 4.6 Low), Copilot from Microsoft (Smart), Gemini from Google (Gemini 3.5 Flash, Thinking level Standard), and Perplexity from Perplexity AI (Sonar 2). We used the free versions of each AI tool with their default settings and LLMs. Each AI tool received the same prompt with no additional follow up, and we edited their full outputs for length and clarity.
(You’ll notice that this time we only used five AI tools, not seven like we did in our previous article. Duck.ai and Lumo produced limited or explicitly uncertain results in our first exercise, so this time we focused on the tools most likely to generate more accurate responses.)
Before landing on our final prompt, we also tested two variations to see how the responses changed with different framing. Our first attempt excluded new seasons of existing shows entirely, and the responses were notably thinner, with some AI tools explicitly flagging how difficult it was to find podcasts that fit our criteria. For our second attempt, we added specific podcast titles directly into the prompt. The result? The responses mirrored those shows right back to us, treating them as an example of what we wanted to see. The lesson here is that even minor prompt changes, and especially the introduction of specific examples, can meaningfully shape AI tool responses. That’s a useful thing for podcast marketers and researchers to understand before treating any single set of AI tool results as a definitive response.
Why does this exercise matter?
While we used the same AI tools in this exercise, the LLMs powering them have changed since February, and that matters more than it might seem. Models are updated, retrained, and replaced, sometimes on a rolling basis and sometimes with notable version jumps. The five tools we used in this exercise are operating on more recent and capable models, which means they have a broader and more current base of information to draw from when answering questions. Comparing results across a span of several months can be the difference between surfacing a podcast that launched in early 2026 versus having no record of it at all. From our standpoint, that’s a significant difference.
There’s also a discoverability angle here. When a brand manager or a journalist or just a curious listener wants to know what podcasts are worth listening to, they’re most likely starting with an AI tool instead of a search engine. Even Google has made their “AI Mode” the default behavior for their flagship search feature. An AI tool gives people a curated, conversational answer that carries implicit authority. Understanding which shows are appearing in AI tool responses right now serves as a form of competitive research, and for podcasters thinking about their own discoverability, this can help show a clear picture of what building a robust online footprint is working toward. In other words, the outlets or sources that are pulled into AI should be the ones you incorporate into your podcast media publicity strategy for earned coverage and features.
Like our previous AI exercise, we did not alter the responses we received. Any “hallucinations” in these responses are something that podcast creators, marketers, and strategists should be aware of as they continue to engage with AI tools for this type of research.
To date, what are the best podcasts of 2026 according to AI tools?
Here is the prompt we used:
You are a podcast enthusiast who always knows about the best and newest shows.
I want you to find the top podcasts of 2026. These must be podcasts which fulfill the following criteria:
- The podcast began publishing episodes on or after January 1, 2026
- The podcast has at least two press mentions
Again, these should be completely new podcasts which began publishing episodes on or after January 1, 2026. You can also include new seasons of podcasts that began publishing episodes in 2026, but the primary focus of your analysis should be for completely new podcasts, so be sure to reflect that in your analysis accordingly.
The output should include the following:
- The name of the podcast
- A link to the podcast on Apple Podcasts or Spotify
- A short summary of the podcast – 100 words or less
- Links to articles where the podcasts are mentioned
Let’s take a look at the results we received from each tool.
ChatGPT
“I dug into 2026 launches and applied your filters as strictly as possible.”
- First episodes published on or after January 1, 2026
- Completely new podcasts are prioritized
- At least two independent press mentions or notable media writeups
- Excluded long-running podcasts that merely started a new season in 2026
“One thing worth noting: we’re only halfway through 2026, so the pool of brand-new shows with multiple press mentions is still relatively small.”
“If I were ranking the most notable brand-new podcast launches of 2026 so far based on critical attention and media coverage, I’d put them in roughly this order.”
“The first three have received the strongest combination of editorial attention and recommendation-list placement, which tends to be the best early indicator that a new show is breaking through beyond its built-in audience.”
Claude
“Here is the full, carefully researched breakdown, with the primary focus on completely new podcasts and one clearly labeled new season.”
Completely New Podcasts
- City of Lights
- Money Trauma
- The Secret World of Roald Dahl
- Pop Syllabus
- Five Miles From Home
Editor’s Note: Five Miles From Home premiered on June 8, 2026. We prompted Claude on June 2, 2026, though two episodes were available to Dateline Premium subscribers. This information surfaced from a press release about the podcast.
New Season (Notable Inclusion)
- Creation Myth (new season of CBC’s Personally)
Copilot
“Below are the top new podcasts of 2026 that (1) began publishing on or after January 1, 2026, (2) have at least two press mentions, and (3) prioritize completely new shows, with a secondary section for new 2026 seasons.”
Top New Podcasts of 2026 (Fully New Shows)
“These shows launched for the first time in 2026 and have multiple press mentions.”
- Creation Myth
- Money Trauma
- City of LIghts
- Love + Radio: Blood Memory
New 2026 Seasons (Secondary Category)
“These are not new podcasts, but new seasons that began publishing in 2026 and received multiple press mentions. You asked for these as a secondary focus.”
- World of Secrets: The Darkest Web — New 2026 Season
Gemini
“The narrative landscape in audio has seen some incredibly focused, independent documentary drops this year. Two standout selections perfectly align with your updated criteria, both debuting in early 2026 as completely fresh concepts that have since captured critical media attention.”
- City of Lights
- Money Trauma
Perplexity
“Here’s a refreshed, stricter shortlist focused primarily on completely new podcasts that began publishing in 2026, with a small secondary set of 2026-starting returns/new seasons only where they also fit your date rule. All included shows have at least two press mentions in the sources I found.”
Completely new podcasts
- Money Trauma
- City of Lights
- Pop Syllabus
2026-starting newer seasons
“These are included only as secondary options because your updated prompt allows new seasons that began in 2026, though the main focus remains on fully new launches.”
What should podcasters take away from this exercise?
Since we were only looking at the first half of 2026, it makes sense that the same few shows appeared in responses across nearly every AI tool. City of Lights and Money Trauma, sometimes both podcasts, appeared in every single tool’s response. That level of consistency is meaningful, and suggests that those shows have established the type of online footprint that LLMs recognize, cite, and surface in AI tool results.
Also, different AI tools still give very distinct answers. Just like in our February exercise, ChatGPT and Copilot gave more structured, neutral responses. Gemini was more opinionated, focusing on independent documentary work over other types of podcasts. Claude was thorough and careful about distinguishing between new shows and shows with new seasons. Perplexity leaned on more explicit citation sourcing than the other AI tools. These differences mean that the same prompt can produce different responses based on which AI tool you use for research.
Even with the variety of responses, there were a few consistent takeaways that we uncovered:
- Podcasts with mainstream media outlet backing still travel further in AI results than podcasts without them. Shows with network affiliations, major outlet partnerships, or from established production companies surfaced more reliably in responses from AI tools. That’s consistent with how LLM training data is weighted: earned media generates more text, in more places, with a greater frequency than coverage of independent work. Creators with fewer PR resources might face a steeper climb towards AI tool visibility, even if their work is critically acclaimed.
- On the other hand, press coverage from podcast-specific outlets matter a lot. Podcast Review appeared as a citation source in every AI tool’s response, namely their article “The Best New Podcasts of 2026 So Far.” For podcasters thinking about press strategy, this is why pursuing reviews in podcast-specific trade outlets is important. Check out our article (produced in collaboration with Podcast Review) about how to get your podcast reviewed by the media for a deeper dive.
- The gaps are still the opportunity. The same demographic and genre blind spots that we talked about in our February 2026 exercise are present here. Comedy, fiction, sports, religious programming, non-English-language podcasts, and shows made by and for underrepresented communities are largely absent from responses across all the AI tools we used. For marketers and strategists, the voices and types of podcasts missing from these results are in underserved territory that AI tools may one day account for as their training data updates and broadens.
Conclusion
As with our earlier exercise, these results should not be treated as authoritative even with the timeline we specified. The podcasts that appeared in these responses are ones with a combination of press coverage and a strong online footprint. It’s a limited snapshot of how AI tools see the podcast space right now, and knowing how that snapshot gets built is the first step towards shaping what future responses could look like. We will continue these exercises as these AI tools update.
Interested in building the kind of presence for your podcast that shows up in AI research results? Learn more about our podcast growth services or subscribe to our free newsletter for analysis like this delivered to your inbox.
