A few weeks ago I wrote a blog post exploring generative artificial intelligence and audio generation using platforms like ElevenLabs and Stable Audio to create synthetic voice and sound effects for a totally AI generated podcast.
This morning I saw a video from educator and YouTuber Alex Gray where he reflected on a LinkedIn post from US academic and AI expert Jason Gulya on using ChatGPT to create podcasts, and it got me thinking of the connection between the synthetic podcast I’ve created using ElevenLabs and the new features of ChatGPT.
So here’s my reaction to Alex’s reaction to Jason’s reaction to ChatGPT’s voice mode…
In this post I’m going to show you both the synthetic podcast and the process of creating it from start to finish. But I want to leave you with the question of whether this is something we really want, whether this is an inevitable progression of these technologies, and where your personal line is between how much human generated and how much synthetic media we want in our lives
First of all, here’s is the mini podcast episode:
The process
Like Alex’s video my process was ridiculously straightforward. Here it is:
- I created a fake podcast transcript with a very brief prompt,
- ChatGPT separated out the speakers
- I created the audio for my speaking part using my ElevenLabs synthetic voice
- I recorded ChatGPT’s voice for Fakey McFake. I could of course have used a different ElevenLabs model but I wanted to demonstrate the other app.
- I dropped the audio into Adobe Premiere Pro and used the AI transcription to generate an editable transcript.
- Drag and drop to edit the speaking parts into the right order (editing directly from the transcript)
- Add the music already generated by stable audio for the previous blog post
- Export the mp3
It’s obviously too short to constitute a real podcast, and it’s not going to win any awards in terms of originality or the quality of the delivery. But as I keep saying, we are on day one of these technologies and it’s almost inevitable that this is the trajectory of generative AI, at least in the short to mid term while as a society we come to terms with the implications of generative artificial intelligence, deep fakes and synthetic media.
Whether or not you enjoyed listening to the podcast, I hope you’ve enjoyed this post and the conversation starter about what the future might hold for good and for bad.
Like or share the post with the buttons at the bottom of this page, and if you’d like to get in touch, contact me using the form below:
Leave a Reply