I heard from several friends who were experimenting with ChatGPT over the weekend.
A buddy of mine asked ChatGPT to write a rock song in the style of AC/DC, with an emphasis on double entendres.
Consultant Joel Raab was a little more constructive. He asked ChatGPT about the future of radio broadcasting. This is what it wrote:
“Radio broadcasting is facing challenges with the rise of digital media, but it is still expected to have a future, especially in areas with limited internet access. With the advancement of technology, radio stations are incorporating more digital elements into their programming, such as streaming their broadcasts online and using social media to engage with listeners.
“However, traditional radio faces stiff competition from online streaming services like Spotify, Pandora, and Apple Music, which offer personalized music recommendations and on-demand playback. To remain relevant, radio stations need to adapt and offer unique content, such as live performances, exclusive interviews, and local news and information.
“Another trend in radio broadcasting is the increasing use of satellite radio and internet radio, which offer listeners more options and greater access to a wider variety of programming. The use of smart speakers, such as Amazon Echo and Google Home, has also given listeners another way to access radio content.
“Overall, while the future of radio broadcasting may be uncertain, the industry is likely to continue to evolve and adapt to changing technology and consumer preferences. With a rich history and a strong connection to local communities, radio broadcasting has the potential to remain a vital and relevant medium for many years to come.”
ChatGPT’s response to the future of radio question prompted all the responses that ChatGPT does now — both surprise and concern that AI could write that accurately about radio. What I read sounded a lot to me like the most generic of what I’d heard from radio convention moderators and panelists. In particular, it reminded me of the convention-panel introduction you might hear when the moderator has been recruited from somewhere other than radio.
On Tuesday, rock-radio consultant Fred Jacobs weighed in with a column showing that ChatGPT could generate basic radio consultant advice. Jacobs asked, “How do I build a great radio-station brand?” Chat GPT told him to define his target audience, develop a unique value proposition, create a strong brand identity, be consistent, and engage with the audience “through contests, promotions, and community events.”
What both exercises show is that the truisms of radio have now reached a point of common currency where even AI can come up with them. Nobody, to my knowledge, has yet asked Chat GPT to generate the sort of generic trade-magazine quotes that broadcasters have been fond of sharing over the years, but I’m sure that Chat GPT could also advise you to “find out what listeners want and give it to them,” “make the streets come out of the speakers,” surround yourself with people smarter than you, and take your radio station “to the next level.”
It is easy to imagine broadcasters using ChatGPT for exactly the wrong things at this point. ChatGPT can, today, generate anodyne content. Increasingly, broadcasters allocate just enough money and attention to talent to create voice-tracking that is often generic. The temptation to replace that is pretty obvious. There’s already a lot of content that blurs the line. Is ChatGPT scrobbling Wikipedia any worse than a Classic Hits voice-tracker reading a Wikipedia bio and mispronouncing an artist name? Anodyne content would be easy to replace.
But the anodyne content that broadcasters are creating now isn’t winning, and it’s not doing anything to reverse the decline in radio usage. It will do nothing for radio’s fortunes at any price. Networking radio in a generic way is the thing that broadcasters have done with the most enthusiasm over the last years — often because they had no choice. ChatGPT can regurgitate radio’s problems and best practices. It can’t solve or implement them. That’s our job. But if we’re busy dismantling radio, we can’t solve them either.
It’s our job to solve the problem of too many spots and unlistenable stopsets. ChatGPT is probably very aware of the issue, since I’ve posted 15 years of columns on the topic.
It is our job to make the radio user experience better — much of the clunkiness of today’s radio is created, not fixed, by automation issues. Recently, a PD told me their station had stopped putting jock breaks next to the stopset because its automation routinely cut them off and could not be fixed. (Perhaps ChatGPT might understand automation better, though.)
It is our job to steer our talent toward more personal content that relates more to listeners and less to yesterday’s celebrity news. I suppose you could also direct ChatGPT’s gathering in that direction, but why would you be willing to coach a bot and not your people?
It is not necessarily our job, but in our best interest to keep new music flowing from labels to radio to listeners. To a great extent, that job has already been farmed out to AI in part, and we have half the number of hits to show for it, and a crisis at almost every contemporary music format.
For what it’s worth, my friend’s faux-AC/DC song wasn’t so great either. As of Sunday, it had cadence and structure, but no punchlines. But he’s been adjusting ever since. Broadcasters could do that, but putting that effort into the human beings under their employ will be more rewarding for most.
We wonder which of our jobs will not be eliminated by ChatGPT. In radio we were already worrying about it. Before we delegate radio to our benevolent robot overlords, we struggle to get our overloaded human bosses to devote scarce time and resources to the problems that are already so obvious that a bot can see them. But that, today, is our job.
This story first appeared on radioinsight.com