For some podcasters, AI is an exciting new technology they happily embrace. For others, it’s a cause for concern and makes them question their role in their own content. Both are happening simultaneously, and both have valid perspectives.
According to this 2024 study, over 57% of podcasters use AI-powered software to support their workflow. You can find tools that will entirely automate every step of your content’s production, from scripting to editing to promotion. You can even use voice cloning technology to create episodes entirely without a human host.
In an age when most people are rushing to solve every problem with AI, it’s worth pausing to consider how you can use the technology to genuinely improve the quality and relevance of your content. In this article, I’ll explain why technology isn’t a replacement for human strategy and creativity and how you can leverage AI’s strengths to make higher-quality, more sustainable content.
The limitations and risks of AI-generated podcast content.
When you hear AI-generated podcast content, it can be unsettling. As of right now, it’s not hard to tell an AI host apart from a real one, but it will be one day.
The possibility of technology replicating human storytelling poses the question: If AI can articulate your message, why do your listeners need you?
Before diving into AI’s risks and limitations in podcasting, let’s define what I mean by “AI.”
AI tools in podcasting can be anything from editing assistants and transcription services to promotional content generators. But in this section, I’ll focus on AI-generated hosts and voices that aim to replicate and replace human hosts. These tools raise unique concerns for the industry. Later, we’ll explore other AI tools and their role in your content creation process.
The limits: Lack of connection with listeners, ethical concerns, and psychological implications.
As I discussed in this article on the psychology of podcast listening, there are deep emotional reasons that people listen to podcast content. The common thread between them boils down to shared humanity. This short video on why sharing headphones is so romantic does a great job of explaining this.
No matter how good AI is at weaving a story, it is ultimately mimicry and will not be able to replace lived experience and human emotion, or replicate genuine humor or emotional depth. Unfortunately, AI can satisfy a listener’s core psychological need to connect with another human. But it shouldn’t. (Don’t get me started on AI significant others.🤮)
“Should podcasters be quaking in their boots? Probably not. While the AI capability is scary and impressive it likely will never capture the rich sound and rhythm of human voices riffing, trash-talking, debating, and connecting with listeners. AI can’t replicate the authentic spontaneity, personality, and laughter that makes many podcasts engaging and keeps audiences returning for more. That kind of magic is hard to automate. Even at its best, AI content often feels just a bit off.”
–Steve Goldstein, Founder and CEO of Amplifi Media
The Verge recently published this piece on an Australian radio station that used an AI-generated host for months on its show “Workdays with Thy.” The station modeled the host’s voice after one of its real-life employees, but it never told the audience it was AI-generated. Listeners didn’t notice for months.
This example highlights why transparency is critical. If your audience finds out you’ve used AI without telling them, it can damage trust, especially in B2B.
We recently launched “Sandbox: A behind-the-scenes production” as a place to test things. One of those things is an AI-read version of our articles. I start every AI article read with “Hey, this is Jeff Large, founder of Come Alive, and you’re about to hear an AI-read version of…”
– Jeff Large, founder of Come Alive
B2B brands’ reputations and success rely heavily on strong relationships with prospective customers. Most of our clients care about building personal and/or brand authority. Some are leaving a legacy. Others care about building stronger relationships with the relevant people in their lives. AI-generated content isn’t conducive to those goals, and the importance of relationship-building in the B2B space makes it even riskier than for less relational industries.
AI-hosts could work for a company like Dbrand, whose whole operation is run by robots. A completely AI-generated podcast could be funny and on-brand for them. But for B2B podcasters, whose goals are greater recognition and stronger relationships, listeners need to care about them. Not AI.
The risks: Potential for misinformation, bias, and low-quality content.
One of the most significant issues with AI in podcasting is the potential for spreading misinformation. Current AI language models can be convincing and still incorrect. This inaccuracy happens because the language models are trained to predict text patterns; they don’t actually understand the content.
Establishing your brand as a reliable source of information is integral to building trust and getting quality leads and conversions.
If you look at these AI-generated podcasts created using Google’s NotebookLM, you’ll see that a lot of time-intensive prompting is required on the front end to generate a decent product. And even with good input, the tools can still make mistakes. NotebookLM also raises concerns about plagiarism and making it easy to rip off other creators’ content.
For example, these AI hosts unexpectedly acted annoyed when real listeners interrupted them and exhibited behaviors that felt unnatural and adversarial. Google had to train the annoyance out of its software.
To “train” anything out isn’t authentically human, and it’s one of the issues I have with AI editing—the “remove awkward pauses” feature. Sometimes those pauses, the annoyed response, and similar reactions are exactly what make stories and interviews interesting. AI doesn’t currently understand that.
Take StartUp Podcast’s How Not to Pitch a Billionaire episode. The way Alex Blumberg is failing at his pitch is painful to listen to, but that episode set the tone for the first season. It was amazing.
Or the interview with Larry King and Jerry Seinfeld. Larry didn’t do prep work for his interviews intentionally, so as to take a similar position as the listener. At one point of the conversation, Larry states that Seinfeld wasn’t cancelled and in fact Jerry gave up the show. Jerry loses his mind, completely insulted.
Sometimes, awkward pauses and unexpected emotions are good. They’re human. They develop a tension and an element of surprise, a feeling that anything could happen. Training those moments out of podcast content will make episodes sound more polished, but at a cost.
So what’s the answer? I think it’s a balanced approach: using AI to support your content, but not over-relying on it to the point that the human element is lost or trust with your audience is compromised.
3 Ways AI can enhance podcast production.
Despite the risks and limitations, it’s best to view AI as an ally in podcasting. Certain tools can help B2B teams cut costs, streamline the tedious parts of production, and help podcasters publish quality content consistently and with less overwhelm. Let’s look at four ways AI can benefit you.
While AI tools can streamline specific parts of the podcast production process, as a B2B company, you’re likely not handling post-production in-house. You probably outsource post-production to a producer or freelancer who takes care of editing and sound, and they’re the ones using the tools. That said, here are some ways AI has meaningfully supported our team’s workflow and the tools we’ve found most helpful.
- Transcription and show notes: Transcriptions are constantly getting better, and it’s a great place to use AI because it can timestamp and summarize accurately way faster than a human can. My team has used ChatGPT, Capsho and Descript to transcribe episodes and create summaries for the show notes. However, these tools are just a starting point that shortens the time spent doing them from scratch. We always have copywriters and editors take bits and pieces of the AI-generated copy and turn it into polished versions.
- Mixing: AI isn’t great at editing yet, but it is pretty good at mixing. For example, the human producers on my team review most episodes and arrange them, omit sections, record pickups, etc., to make them the best they can be. Our editors and audio engineers do the heavy lifting–editing complex parts, mistakes, and specific timing.
That said, Descript’s Studio Sound feature works well if the source audio is loud enough, and its Compressor and Limiter can make the human voice sound awesome. This feature is really helpful for exporting audio at a consistent level.
- Marketing and promotion: AI’s current strongest feature isn’t choosing what to feature but speeding up the editing and asset creation process. I’ve tested the auto-generated clips, and they kind of suck. In my testing, AI-generated clips hit at about a 1:3 usable-to-garbage ratio. I still rely more heavily on the interviews and producers’ notes to find the best clips, depending on what our goal is.
In short, AI lightens the workload, but it’s our team that infuses it with nuance, style, and
strategy. Now, let’s look at where AI should not replace humans so you know what to look out for if you incorporate it into your workflow.
Where AI should not replace humans.
While AI is a great tool to help with very specific and technical aspects of podcasting, some critical areas still need a human touch to ensure the content resonates with your audience. Here’s where AI falls short:
- Editing for clarity and impact: AI tools can easily remove filler words and silences that would take humans a long time to edit out themselves. But they can’t make the nuanced decisions that skilled editors do. AI can automate some edits, but it can’t fully replicate a skilled producer’s nuanced decisions, like leaving a silence or a sigh or removing some amount of the conversation for clarity or impact. Human editors know when to leave a pause for emotional impact, how to maintain the flow of a conversation, and when to cut for clarity without losing the essence of the discussion. These creative decisions are integral to compelling storytelling and require an intuition and judgment that AI isn’t (currently) good at replicating.
- Strategy and storytelling: AI doesn’t understand the emotional journey of a story or the subtleties of audience engagement, which often make the biggest impact. Crafting a narrative that resonates with listeners requires lived experience, empathy, and vision—uniquely human attributes. Podcasting is not just about words; it’s about smart marketing strategy and meaningful points of connection with your listeners.
- Brand voice and relationship-building: In industries like B2B, building trust and rapport with your audience is crucial. AI can’t mimic the personal connection that human hosts establish with listeners over time. Your audience wants to feel heard and understood, and a genuine, relatable brand voice is key to building that trust. Over-relying on AI-generated content can lead to a disconnect, making your content feel impersonal or mechanical. Plus, most experts want people to grow to trust them, not the AI voice.
- Creative judgment: Overusing AI for content creation can inhibit the critical thinking and creativity that make your podcast stand out. Humans bring unique insights, perspectives, and emotional intelligence to their work. No matter how quickly AI strings sentences together, you still need humans to ask questions, like “Does this content align with our brand? Does it serve the listener’s needs? Is this engaging or impactful?”
It also helps me to remember that AI returns the group’s average. Sometimes, that’s helpful, and sometimes, it’s what it is: average.
AI can be an incredibly helpful tool for speeding up repetitive tasks and supporting your workflow, but it can’t replace the creativity, intuition, and strategic thinking that make podcasts memorable and meaningful. Humans are still essential for crafting content that truly resonates.
Final thoughts.
This is a tricky topic to cover. Some people have already sold their souls to AI, and others hate its digital guts. All I know is that AI is changing things fast, and many of the concerns surrounding the topic are a matter of “when,” not “if.”
I remember a year ago, AI generated this interview between Bill Gates and Socrates. At the time, most of my peers in the podcast space were commenting on how awkward it was and how it didn’t sound human at all.
But it’s only a matter of time before it does. Take the drastic improvements in Midjourney as a visual example of the technology’s progression.
I’m sure most artists, illustrators, and photographers were laughing back in March 2022. They aren’t now.
I think the same will happen to many other professions, including my own.
For now, I view it as a tool, and if I can help me speed up a process effectively (not just efficiently) and be more human, I’ll use it. So the key is to know where we stand now so we can remain true to our values later (when it gets really crazy).
In a world full of noise and copycats, the most powerful thing you can be is real. If you want to learn more about podcasting to amplify your brand’s authentic voice and build trust with your audience, book a free discovery call with me. I’d love to learn more about your business and how podcasting can help you reach your goals.