Radio vs. AI: The War Radio Invented

Many of us remember when the tools in our trade were quite different than they are today. For example, I remember using index cards to schedule music. It was a time-consuming and painstaking process. When the first music scheduling software systems became available, they offered tools we could never get from the index cards.

And then, almost overnight, those early scheduling programs started doing things we could only try to do with paper—enforcing separations, keeping rotations consistent, and letting you adjust a log in seconds instead of spending half a day reshuffling cards. It wasn’t just faster… it was cleaner, tighter, and way more reliable—because the system was catching the little things that used to slip through when you were tired, rushed, or just trying to get the day done.

I can remember how difficult audio production was back in the day—multi-track machines, razor blades, and the kind of edits that took more courage than skill. I could hear the perfect spot in my head, but getting it to land that way in real life was another story. Most of the time, the finished product came close… but it rarely matched what I imagined.

Even when we moved into digital production, it wasn’t like the problems disappeared. The tools got better, but the gap between “what you want it to sound like” and “what you can actually pull off” was still real.

And dialogue spots? Those were always a challenge. You’d line up a few people around the office who were willing to help, but they didn’t have the polish, the timing, or the delivery to make it sound truly professional. Then the layoffs started, and even the stations that did have strong voices in the building suddenly didn’t. I remember local stations running whole sets of commercials where everything was the same two voices… and sometimes it was just one.

With each technological breakthrough, radio—like almost every other industry—benefited. Some of the advancements simply saved time. Others made the final product infinitely better and saved time at the same time. And yet, no matter how obvious the improvement was, there were always a few people who resisted it.

In my experience, the biggest holdouts were usually the ones who had become really good at doing things the “old way” before the newer tools showed up. They weren’t just protecting a workflow—they were protecting an identity. When your unique skillset becomes something anyone can learn faster, it’s easy to feel like your value is being threatened… and like your earning potential is about to get diluted.

Today, a new wave of tools—this time fueled by AI—is inspiring a backlash. iHeartMedia even rolled out a “Guaranteed Human” policy that draws a hard line against AI-generated on-air personalities (and even AI music with synthetic vocalists), like that was ever the actual aspiration of this tech in the first place. A lot of professional voiceover talents are resisting too, because AI has the potential to generate imaging and production voices that are freakishly good—good enough to make people nervous.

I’d urge everyone to do what I’ve done with this technology: take a breath and assess where it fits the same role other major breakthroughs played for us—music scheduling software replacing index cards, and digital tools replacing razor blades. The best advancements didn’t “replace radio.” They made our workflows smarter, faster, and more consistent—and they freed us up to spend more time on the parts of the job that actually require taste, judgment, and real creativity.

Smart voiceover talents should be looking at licensing structures that let stations create approved, contract-based “voice likeness” models—so clients can use AI to generate certain kinds of work independently, without cutting the talent out. As the technology keeps improving, those models could produce promos and spots with the same inflection, nuance, and realism that made the original voice valuable in the first place. The difference is, instead of fighting that future, voice talent can help define it: clear permissions, clear boundaries, fair pricing, and a system where their voice becomes a scalable asset—not something that gets copied for free.

This may sound like getting ahead of the technology, but imagine what it could mean for voice talent: getting paid for a God-given gift without spending endless hours in a production room doing take after take—because the client can audition options, find the exact read they want, and land on a final version without burning up studio time. Done right, it doesn’t reduce the value of the voice… it protects it, packages it, and makes it easier to use while the talent still gets compensated.

Talk about localizing a syndicated show. What if you could produce fresh liners every day using the voices of your syndicated talent—tight, clean, and specific to your market? AI could make it possible to write and generate basic local liners daily: station name, city mentions, weather teases, contest tags, sponsor reads, even quick “coming up next” lines that sound like they were cut in-studio.

And it wouldn’t have to be the Wild West. You could build it with guardrails—host-approved language, strict copy templates, and an approval step before anything ever becomes audio. In other words: the market gets stronger localization, the show stays on-brand, and the talent stays protected because nothing gets generated without control and permission.

Another tool in the arsenal is jingle production. Writing the lyrics and uploading a simple, sung acapella melody can now yield professional jingles that can rival what used to require a high-priced production house. And never mind the objection that “if everyone can do it, then it has no value.” Everyone has access to canvas and paint, but not everyone can create the Mona Lisa.

At Sonic Attention, I’m able to use real songwriting talent—original lyrics, original melody, and real creative judgment—to deliver something far beyond “click-and-create” output. The tools might be newer, but the part that matters most is still human: the hook, the phrasing, the storytelling, and the instinct for what will actually stick in someone’s head.

For one thing, my melodies and lyrics are copyrightable because they’re human-authored, unlike works that are generated entirely by AI with no meaningful human creation behind them.S

This new wave of AI tools isn’t some alien invasion—it’s the latest version of the same kind of leap we’ve seen before. Used the right way, it can revolutionize workflow, restore polish and consistency, and help radio claw back some of the quality that endless downsizing has stripped away. And instead of replacing humans, it might be the very thing that finally gives radio enough lift to grow again—because when you can do more with less without sounding like less, the business has room to breathe.

And the best part? This time, nobody has to bleed on the studio floor to make it happen… because we finally upgraded from razor blades to results.