Why AI Will Disrupt the Audio Landscape
By Sonia Yee
If you run a business then ChatGPT and AI are most likely front of mind. And equally, if you’re working for an organisation, regardless of the type of industry, there could be a chance that your job, or at least parts of it, will become automated.
Those in the podcast industry should also stop and consider where AI could enhance the production process or detract from it.
Podcasts by their nature are about an intimate listening experience, and a story on Inside Radio suggests AI will provide a solution - if we don’t like the presenter’s voice, we can switch it out for AI voices, ‘Leo or Luna’, instead.
But will AI stand a chance in an environment where the uniqueness of the human voice has laid the foundation for the podcast medium in the first place?
The argument in support of AI hosts ‘Leo and Luna’ taking over the mic is about having a ‘personalised listening experience’.
Don’t get me wrong, but isn’t that like deciding you want all of your favourite songs sung by one artist across every single genre - only worse - because it’s being generated by a machine?
Artists pick a genre because it suits the timbre of their voice. While some musicians cross into different styles of music, there will be some areas that are outside of their vocal comfort zone. Think, Mariah Carey dropping some rhymes on a hardcore rap album or screaming into a microphone for a death metal track.
What I’m getting at here is that podcasts already provide a ‘personalised listening’ experience because we have so many options to choose from. To adopt AI as a ‘solution’ because we’re annoyed by the presenter’s voice makes for grounds to continue on our search for what we do want to listen to - and the wealth of audio on offer makes that possible.
But that’s also the problem; wading through the pure abundance of podcasts can be a time-consuming experience, if you want to find the golden egg.
Remember too, that we don’t have to like everything, and we never will. Nor should we accept that AI needs to become part of an industry where storytelling and craft should be at the heart of it.
There is nothing like listening to a host whose voice can take you somewhere you never imagined, including opening you up to a new way of seeing the world.
I personally, want that human connection in an audio experience; one that is rich, expressive and full of emotion. AI is unlikely to offer the nuance and will surely be no better than a ‘wooden’ presenter whose performance has already driven you to switch to Netflix.
My question is: why should we try to solve a problem that doesn’t exist?
If anything, we could run into the danger of repetition and uniformity in the audio landscape, which is already ever-present in the podcast space. The ‘one-on-one, deep dive interview’ has been reproduced and replicated a million times over, and is largely remiss of any creativity, innovation or inspiration.
This format, layered with a machine-generated voice that is impossible to connect with, will only create distance between the listener and the content, and that is the inverse of what you want.
Technology is being used to solve some of the world’s biggest problems, but so too, it can create them.
Yes, there are likely to be other ways of integrating AI into the podcast workflow in a much more productive and clever way, but this is not it.
Whatever the case, Leo and Luna will have to work pretty hard to gain traction, especially for those of us who love the audio craft - and a listening experience that reinforces the importance of the human experience, and our deep desire to connect.