The Ghost of Podcast Present: Using AI in Your Show
Human beings have more or less been obsessed with robots taking our jobs since the term “Robot” was first coined in 1920, in a play that depicts just that!
Then, of-course, there was I, Robot, where in 1950, Isaac Asimov lays out his “Three Laws of Robotics.”
Perhaps one of the most telling early work is Kurt Vonnegut’s Player Piano, published in 1952 about a dystopian future where virtually all jobs are replaced by automation.
For something more contemporary (and more optimistic), you can look to John Danaher’s Automation and Utopia: Human Flourishing in a World without Work, which argues that a workless, automated society may allow us to focus on more, “creative, intellectual, and social pursuits.”
You know — those things that actually make us human, and not just being striving toward complete and total efficiency.
No matter what you think, the Generative AI boom of the past year has accelerated conversations like this. Will AI replace me? What can’t AI replace? How can I leverage AI for my job? What’s overhyped?
As a podcaster, you might be wondering these same things.
After all, you’ve likely seen some AI generated content to make you think that human podcasters are on the way out.
So I thought for the penultimate article in this year’s podcast advent, we’d explore the ghost of podcast present: AI.
There will be lots of articles recapping “this year in podcasting,” but if I had to pick a topic that affected the space most this year, it would be AI.
You can think of it as podcasting’s [Artificial] Person of the Year.
The “artificial” part is important. Because even though AI has gotten exponentially better this year, there’s something AI can’t replace: the human element.
What AI can do for us is assist. It can help us generate ideas, research topics, create assets, repurpose content, and make our shows more accessible.
So let’s take a look at how AI is impacting podcasting now — and prepare us for our final ghost, and article of this advent season, with tomorrow’s Ghost of Podcast Future.
Note: I’m going to assume at this point, that you know what AI is. If not, I wrote a handy primer for you.
For more writing like this, subscribe to my newsletter. You can also follow me on LinkedIn
Generative AI is Not Automation
I think the first mindset shift we need to understand is Generative AI is not automation — that is, it doesn’t necessarily do work for us.
It can’t (or shouldn’t) write scripts, do unchecked research, or record for us, no matter how convincing some think it is.
Instead, we need to think of it as an assistant — something that helps us do our jobs better.
As long as we keep this in mind, we’ll be able to leverage generative AI properly and still make compellingly human content.
Let’s dive in!
Planning and Research
Sabermetrics transformed baseball.
A sport that’s so data-heavy used to be based on gut feeling and incomplete scouting reports — for some players, it still is.
But for most players, Sabermetrics gives them a plan. Pitchers know what pitches are most effective against specific batters. Batters know what pitch they’re most likely to see in a 2–1 count. Fielders know what part of the field a batter is most likely to hit the ball to.
In other words, they have a much better plan, thanks to the vast amount of data they have access to.
Generative AI is similar — it gives us access to a vast amount of data in an incredibly accessible way.
We can leverage tools like ChatGPT, Google’s Bard, and Bing Chat to organize and surface basically anything we want to learn more about.
As a result, when you plan your podcast, Generative AI can give you incredible insight into a number of areas.
- What people are talking and writing about
- What questions people are asking
- What is already being done
…and much more.
One of the most powerful uses I’ve see here comes in the form of defining your ideal audience.
Listener Research
Something every podcaster should do is define a clear audience — they can do this through creating a mission statement for their podcast.
The mission statement defines who they serve, what problem they help their listeners solve, and how they solve that problem.
Knowing this, you can do some listener research by:
- Describing your audience to the AI and having it come up with 1–2 listener avatars.
- Asking the AI what questions your potential audience has about your podcast’s topic.
- Learning where you can find your audience online
Beware though! Just like Sabermetrics can give hitters a bunch of data, it can’t swing the bat for them.
AI can help you do the work. It can’t do the work for you.
Better Guest and Topic Research, too!
You wouldn’t say a bat hit a home run, right? A baseball player still needs to swing it at the perfect time to launch that thing over the wall.
Just like with pre-launch planning, AI can help us on a per-episode basis.
Leveraging AI to help us with guest and topic research has gotten so much better this year. We can now use ChatGPT and Bard to get actual web results, which means we can source and validate research more easily.
We can also use it to help us come up with better interview questions — either by surfacing topics we would not have thought of, or by summarizing recent interviews from the same guests and avoiding those topics during our interview.
This sort of better preparation will all you to differentiate yourself through great content — something that we learned from the ghost of podcast past.
Plus, we can spend some more time on the actual content, thanks to the transformative nature of AI on the production side of the house.
Production and Post Production
I think the biggest strides on the actual production side is the permeation of podcast summarization through transcription.
Less than 2 years ago, I was still making the case that podcasts need transcripts. Today, thanks to AI, they are a crucial part of the modern podcast workflow.
I don’t think I need to convince anyone today that transcripts are invaluable for both podcaster and listener alike.
Not only have AI transcription services gotten more accurate (though definitely not perfect), but their utility has grown by leaps and bounds.
Descript makes it easy to edit your podcast by editing the transcript. Both Descript and Riverside leverage transcripts to find and create clips you can share on social media.
Riverside even employs AI through the use of transcription to create a video that automatically switches to whoever is speaking at that moment.
And then there’s the increasing number of summarization tools. From Castmagic to Capsho and SwellAI, to countless others, these tools will take a transcript, process it, summarize it, and give you reusable assets like social media posts, blog posts, newsletter text, titles, descriptions, and more.
Where many podcasters struggle to write compelling copy for their titles and descriptions, these AI tools can provide a great starting point.
That said, AI can’t help us with everything — and that’s an important lesson too.
Where AI Falls Down: Promotion
The number one question I get from people who feel they already have a great production workflow in place is this:
How can I automate my podcast promotion process?
Sadly, there was nary a time I got that question where I was able to come up with a satisfactory answer.
Sure, AI tools can help us “find” and create clips. It can come up with tweets and LinkedIn posts and newsletter copy.
But it can’t possibly know what content is most interesting to our niche audience. It can’t know if the clip it selected falls in line with our mission statement.
It can’t meaningfully engage with our audience online or in communities.
And it can’t force guests to share our show.
And with the ever-changing landscape of social media algorithms, we can’t necessarily rely on those for growth anyway.
Instead, when it comes to promotion, that’s still on us. We can use AI to help us create assets, but we can’t outsource the actual promotion to an AI because it still requires that thing I mentioned at the top of the article: the human element.
And that’s really the purpose of Scrooge meeting the Ghost of Christmas Present, right? To learn about the joy of spending time with family. To see people struggling who are still happy to be together.
To inspire change through common stories highlighting our humanity.
So while AI has definitely shown us it can have a huge impact on our podcast process from planning to promotion, there’s a much bigger lesson we can learn from the Ghost of Podcast Present.
Podcasting still requires something even more important than efficiency: humanity.
For more writing like this, subscribe to my newsletter. You can also follow me on LinkedIn
