(Necessary disclaimer: opinions are my own and not my employer’s or anyone else’s, not now and not ever.)

Do I need to explain what radio voicetracking is? That’s when a jock records their talk bits hours or days ahead of when the bits will actually air. Hit the record button, listen to the last few seconds of whatever precedes our track, say what we’re going to say when the little light goes on, then hit another button when we’re done talking to trigger the next element. By not having to sit through the songs and the commercials, it is possible to track a show that’s several hours long in vastly less time. It is also possible for voicetrackers to be heard on stations in markets far distant from where they live.

A hill on which I would die is that doing a radio show is crafting, like building a birdhouse or throwing a pot. So I prepare a tracked show just like a live one. Whether it’s just one hour or several, I script everything in advance. I have recommended this approach to other jocks, but they tell me it would take them too much time. (I always wonder why a person would go into radio if not to spend time doing radio things, but I guess that’s none of my business.) A voicetracker can walk into a studio and bang out a five-hour show in a relatively few minutes, but it’s gonna be mostly rote DJ stuff, which does little to really command a listener’s attention. Like any other craft, the result you get out of your work is proportional to the time you put into it.

I am forever concerned that my tracked shows don’t sound the same as my live ones. Doing a show in real time means that my talk breaks have a more lively and positive energy than voicetracked breaks, which tend to be more sterile. I am not actor enough to fake that energy to my own satisfaction, although honesty compels me to report that I don’t know if my listeners perceive a difference.

People want to debate whether voicetracking is good, but that ship sailed years ago. What we ought to be debating right now is AI.

Last week, Fred Jacobs published a guest post at his site from veteran radio executive Tom Langmyer, talking about what artificial intelligence might mean for radio, especially as it relates to voicetrackers. One point I had not considered is the ability of an AI “voicetracker” to respond to events on the fly. Langmyer tells about one radio station group that pulled all of the Gordon Lightfoot songs it had scheduled on the weekend after his death because the voicetracking was already done, and it would sound weird to play a Lightfoot song without mentioning that he had died. (At least they cared enough to do that much.) An AI “voicetracker” could have accounted for this, and could certainly update other stuff in real time as well. Certainly this would be a vast improvement over the by-necessity-generic nature of conventional voicetracks.

Langmyer doesn’t mention the very first thing I thought of, however: if an AI “voicetracker” is good, why not replace all of the jocks with AI, which will work for free (after the initial expense), never take vacation or sick days, and won’t bitch about anything ever?

The promise of voicetracking was that a station in West Overshoe could sound like it had major-market talent, which did not turn out to be true everywhere. AI offers a similar promise, only at a more sophisticated level. Will it deliver? I don’t know. Will stations—especially the major chains, drowning in debt with worthless stock—embrace it regardless of whether it delivers what it promises, because of the cost savings?

Yeah, about five minutes after it becomes practical.

Afterword: I have watched the rise of ChatGPT and similar large language AI models with absolute horror. We are not remotely ready for the implications of them. Corporations and entrepreneurs alike are forging ahead with such applications in search of financial windfalls without giving a single microscopic damn about the havoc it is likely to cause. Even if you aren’t concerned about rogue AI exterminating humanity, you ought to have practical and ethical concerns: about the proliferation of deepfakes, about plagiarism, and about whether we really need to hear the Beatles doing Kanye West songs or some shit.

This stuff is world-changing technology on the scale of the wheel and the light bulb, and we’re treating it like it’s Candy Crush on our cell phones.