It took a 40 year vet to solve the AI audio slop problem.

I recently declared that the human ear is a soul detector. Lock eyes, lock ears, same biology. I said AI voices are technically perfect and emotionally vacant. The lights are on, but nobody's home.

I stand by every word.

Then the industry did what the industry always does. It moved.

The newest wave of AI voice platforms isn't building robots from scratch anymore. They're cloning real, working voice actors, with consent, with contracts, with royalties paid every time the voice is used. At the forefront, BionicVO, comes from Dan Price of Oink Ink Radio. Full disclosure: Dan is a friend of mine for decades, and I have profound respect for him as an audio architect. Which is exactly why his premise stopped me cold: AI voices don't sound bad because the technology is bad. They sound bad because nobody is directing the performance.

When a guy who spent forty years directing the best voices in the business says that, I don't hear a tech pitch. I hear a Radio truth. And knowing Dan, I promise you he didn't build this to replace the craft. He built it because, like myself, he too couldn't stand listening to what happens without the craft.

We Knew This All Along

Think about the best spot ever produced. Was it great because the microphone was expensive? No. It was great because someone in the booth said "one more take, but this time like you're telling your best friend a secret." Direction. Performance. Craft. The stuff our business was built on before anyone ever typed a prompt.

So here's the question keeping me up at night: if the source is a real human soul, and a real human director shaped the performance, and the actor gets paid every time... is the ear still fooled? Or is there finally somebody home?

My Honest Answer: I Don't Know Yet

And I'm suspicious of anyone who claims they do. Yes, even my old friend. The purist in me says a clone is a photograph of a soul, not the soul itself. Beautiful, maybe. Alive, no. The pragmatist in me says our stations are already stretched thin, production deadlines are brutal, and a well-directed clone of a great actor beats a badly recorded human reading copy like a hostage statement.

Both of me can't be right. Or maybe both of me can.

 What This Means for Our Business

Here's what I do know:

Consent is the new compression. The platforms that pay actors per use are betting that ethics is a feature. Good. Radio should cheer for that model and be loud about it, because our credibility with listeners is the whole store.

Direction is the differentiator. If AI performance quality comes down to who's directing, then the people best positioned to win are... us. Radio has the deepest bench of performance directors on Earth. We just never called it that.

The live human is still our moat. No clone does a school closing, a snow day, a hurricane, or a hug through the speakers at 6 a.m. That was true last year. It's true now.

So no, I'm not surrendering my soul theory. I'm updating the case file. The ear is still the judge. But for the first time, the defense has hired a very good lawyer. One I've known for decades. And he's annoyingly persuasive.

Keep listening. I know I will.