A couple years back at CRS, I gave a speech about how technology is changing the radio broadcasting industry. To kick off that presentation, I played the AI voice of Taylor Swift greeting attendees, making fun of me, and introducing many in that hotel ballroom to the potentials of Artificial Technology—the good and the bad.
That may have been the first time I said these words about AI:
“Today is the worst it’s going to be.”
And we now know that a truth about AI is that it literally gets smarter over time, learning as it vacuums up more and more information. And yet, while we know the learning is a constant, somehow AI engines continue to make mistakes.
Just like people. In fact, you could reach many of these same conclusions about humans. Learning is theoretically unlimited for most of us. And there’s no question that even as we experience more of the world and life, we are more than capable of making mistakes—sometimes the same ones again and again.
After all, that’s human nature. And maybe AI nature, too.
That’s how I started thinking about it when I ran across a story this past weekend ranking various AI engines on their respective intelligence. Under the Voronoi umbrella, the Visual Capitalist published a fascinating infographic aptly called “The IQ of AI.”
For their comparison, they ranked 24 leading AI models by their intelligence, using a challenging benchmark—the Mensa Norway Intelligence Quotient (or IQ). Visual Capitalist explains the average human score on this test is in the 90 to 110 range. Any score that breaks 130 is considered genius intelligence.
So, who’s got the smartest AI engine? According to Tracking AI, it turns out to be Open AI and their o3 model. According to the scorekeepers, it hit the 135 level, the only engine that cracks the genius barrier.
The rest of the pack is merely good down to average—and a number of AI platforms rank only in the 60/70 range. Maybe that’s not “imbecile” range, but let’s just say many of the people we come into contact with in life are smarter than these AI laggards.
In fairness, Visual Capitalist lets us know the vision AI models all rank lowest, while the top 10 are text-only models.
Personally, I used Open AI’s Chat GPT most. But out of convenience, I also used Bing’s Copilot which only scores an 86 on the Norwegian Mensa test—OK, but far from great.
And that got me thinking and asking questions.
The one that stimulated my thinking was to interrelate the IQ of AI up against many common professions—including the radio DJ.
In perhaps what could be a summer replacement show for Are You Smarter Than a 5th Grader, I started wondering if, in fact, the average radio broadcaster has more IQ points than many of these AI models.
It’s not hard to find intelligence rankers by profession. Many reveal janitors are the least intelligent of the tested job categories, while those in medical occupations and college professors are at the top of the heap IQ-wise.
Source: Quora
But you can look up and down this list and DJs are MIA.
So, I kept digging and found this IQ hierarchy by college major, thanks to Reddit. But it turned out disappointing as it fails to include studies in “Broadcasting” or “Telecommunications.”
The closest I could get was “Communications” with an average-ish IQ of 111 (highlighted below in orange). But my experience is that “comm classes” never had much to do with day to day jobs in broadcasting or even media.
So I did the logical thing—I consulted Chat GPT and asked its powerful engine to tell me the average intelligence of an on-air radio host. And its response?
Essentially, “It’s complicated.’
Below is the response I received to my query.
I really like this response. It’s nuanced, measured, and takes into account many of the variables we might consider when hiring hosts, presenters, journalists—depending on the type of station we’re talking about.
I also like the way Chat GPT makes the connection with verbal intelligence. The expectation for success on the radio requires what they refer to as “verbal intelligence.” Given that need, our AI engine assumes an IQ in the 100-115 range. And incidentally, even at the midpoint of that intelligence range—107-108—that would tacitly place the average DJ as smarter than the majority of AI models as they exist today.
That’s encouraging, of course. But there’s no way radio DJs can stay ahead of the bots by resting on our laurels. In fact, earlier tech chapters suggest than when radio broadcasters get lazy or even arrogant, they are often easily lapped by the field.
And that’s my fear here. In our last couple AQ studies among American air talent on commercial stations, we’ve asked respondents to rate the most important skills sets required for radio success. And in both studies to date—and the one we recently conducted among Christian music radio talent for CMB—the priority to learn more about AI is at the very bottom of the heap mentioned by only 8% of our talent.
Given the rising value of digital revenue to the radio industry, perhaps the smart move would be for talent to embrace learning about AI—what it can and cannot do and how it might enhance rather than replace jobs.
If past is prologue, when broadcasters look down on new technology, take a “wait and see” attitude, and are loathe to learn about it, much less embrace it, the world passes radio by. That point has been repeatedly proved by streaming, mobile, social, and podcasts, to name a few.
AI has made quantum progress since Chat GPT became widely available, but the technology still has a long way to go.
That leaves the door open to radio broadcasters—yes, even DJs—to take the smart money and make AI part of your ongoing education.
Albert Einstein once noted, “The human spirit must prevail over technology.”
For those of us who make our livings talking into a mic, that’s an encouraging observation. But we can only “future proof” radio when we have a strong understanding of what we’re up against.
Originally published by Jacobs Media
