“Dust on the wind
Boots on the ground
Smoke in the sky
No peace found
Rivers run red
The drums roll slow
Tell me, brother, where do we go?”
Interesting song lyrics, right? If you like the ’70s rock sound, you’ll also find that the music is quite good.
If I didn’t know “The Velvet Sundown” was a band created by artificial intelligence, I might be a fan. Spotify calls them a “synthetic music project … not quite human … not quite machine … The Velvet Sundown lives somewhere in between.”
Haters quickly jumped to comment, but the band continues to rack up streams in the millions.
As I write this, broadcast radio is cautious about giving this “band” airplay. Is it the right approach? While I don’t have an answer, it’s clear that radio stations need someone familiar enough with artificial intelligence to proactively decide how to interact with it, far beyond the use of AI DJs.
Using AI to your advantage
- Research: Stations are sitting on a gold mine of audience data received from listeners via social media platforms, email, voice notes and messaging apps. Convert your data into text, upload to AI and bingo, you’ve got aggregated audience insight. If a messaging app is in your communication loop with listeners (see my previous article about WhatsApp), you can tap phone numbers from this group to ask specific questions about music, personalities or talk show topics.
- Emotional intelligence: Humans who develop this skill are more successful than those who don’t. As ridiculous as it sounds, AI will help us scale emotional intelligence to analyze how listeners feel about what they are hearing. Tools like openSMILE can extract voice attributes which can then be emotionally classified to understand how listeners feel. MarketsandMarkets forecasts the emotion AI market to grow from $2.74 billion in 2024 to $9.01 billion by 2030.
- Show Prep: AI is your expert for coming up with hot topics, themes and even sound design. And after the show, taking the transcript, feeding it into AI and telling it to summarize content will provide an ongoing history to prevent repetition. You can also prompt it to expand themes for future shows.
- AI voices: Whether or not you opt for full-time AI DJs, consider how to use AI voices in supporting roles. For example, AI can voice comments or questions about a topic you want to seed for live callers, or perhaps do your traffic or weather. Ever had a vigorous debate with an AI voice? The controversy alone will generate interest.
- Translation: A friend of mine recently spoke Japanese, sort of. AI can now translate your hosts’ voices into other languages while retaining tone and personality. That version could be run on an HD channel or stream. This could be compelling in a market with a significant second-language audience like Miami or Los Angeles.
Authenticity matters, so AI always needs to be acknowledged to the audience. This is important until it becomes so common that it will go the way of “portions of the day’s programming are reproduced by means of electrical transcriptions or tape recordings.”
It’s difficult to discuss AI integration without acknowledging the fear that it will eliminate more jobs. Radio has been running lean for years: small staffs, long hours, tight budgets. However, another way to look at AI is as a tool for relief and support, especially for those still dedicated to America’s number one audio medium.
Read more about this topic in the 2025 Radio World ebook “AI Revolution in Radio & Streaming.”
[Read More Radio World Stories About Artificial Intelligence]
