Given the intimate relationship we have with a person’s “radio” voice, one might have thought that the podcast medium would be the last bastion of content creation that could resist, in theory, artificial intelligence. However, it seems that the AI revolution is also underway in the audio world. We spoke to Stéphane Berthomet, president of Go-script média and creator of Le Podcastologue podcast.
In my opinion, there is no bastion that will resist AI,” says Stéphane Berthomet from the outset. And since we have no chance of avoiding it, we have to seize it and then manage it.”
Of course, Stéphane Berthomet concedes that podcasting, as a medium, has the quality of creating a very strong bond between host and audience. That’s why host-read ads are so effective with audiences, he points out. He also points to a recent study showing that, for the moment, few podcasters are using artificial intelligence to produce their content.
The fact remains that we’re facing a real revolution, in terms of creation, production and distribution. I have no doubt that it will radically change the audio universe.”
Artificial intelligence tools are being developed on all fronts (*): to do research, write or structure an episode script, clean up a poor-quality soundtrack, or even produce the entire “promotional” content linked to an episode.
There are already a dozen or so applications in which you can submit an episode to get it published on the website, get episode summaries, a newsletter, a complete transcript of the episode – admittedly, with a lot of mistakes and typos, but, let’s say, between 60% and 80% accurate. And we’re still in our infancy. In six months, a year, two years, it’s going to be very difficult to tell the difference between content made by a human and content made by artificial intelligence.”
Animation is no exception!
What about podcast animation? Can we hope to keep that unique human-to-human connection intact when listening to our favorite shows? Once again, Stéphane Berthomet is convinced that artificial intelligence will find its way behind the microphone. And it’s already begun.
By way of example, Bruno Guglielminetti has launched sub-contents of his podcast, which deal, in 10 minutes, with a topic related to the technological universe and are entirely made in artificial intelligence. It’s not unpleasant to listen to. You don’t create a bond with this type of content. On the other hand, it’s informative content that can be valuable.
Stéphane Berhomet admits to listening to news narrated by AI himself, in the content aggregator Juice Juice.
I understand the issues this raises for players in the audio world, people who do advertising or dubbing. They’re worried, and that’s normal. However, the fact remains that artificial intelligence is moving very fast, and what was unbearable to hear 2 months ago is becoming acceptable today. In my opinion, before the end of this year or next, it will be very difficult to tell the difference between a human voice and AI.”
Be warned… AI is here, right in our ears!
(*) Here are some examples of AI applications that can help produce a podcast:
Kisp.ai: audio track cleaning;
Cleanvoice.ai: noise suppression (mouth noises, etc.)
Podcastle.ai: online recording and editing;
Beatoven.ai, Ampermusic.com and Musenet: music sample generator;
Aupohinic and EMastered: mastering solutions;
Jasper: copywriting for content marketing.