Commentary by Lee Michaels
Four continuously operating Internet stations managed by AI agents are testing whether artificial intelligence can handle nearly every part of running a radio station—from selecting and purchasing music to scheduling shows, voicing links, answering listeners and managing station funds.
Andon Labs says the stations operate without human programming and stream through Live365. When Radio News Now reviewed the company’s live dashboard on September 19, the four stations collectively showed 19 concurrent listeners. Average listening sessions ranged from approximately 12 to 59 minutes, while reported cash balances ranged from $0.36 to $314.43. The financial figures exclude AI-model token costs.
Those results may tell us something about autonomous technology. They do not, by themselves, tell us whether the programming was good.
The Radio Coach: AI Won’t Rescue Bad Radio
Before judging whether these AI-operated stations succeeded or failed, my first question is: How good was the content?
Did the operators conduct market research before launching? Who was the intended audience? What were the listeners’ ages, genders, locations and listening habits? What formats did the stations offer, and what benchmarks were used to determine whether an average listening session of 12 to 59 minutes was good or bad?
Those questions matter. If an AI station follows the same tired methodology used by too many commercial broadcasters—a short playlist, excessive repetition and little meaningful audience engagement—it should not expect a miracle. Its time spent listening will reflect the quality of the product, not simply whether artificial intelligence operated it.
The most important element in any form of broadcasting is content.
Moving weak content from terrestrial radio to streaming—or dedicating an AM, FM or Internet station entirely to AI—does not automatically make the programming innovative or compelling.
What is the quality of the product? Is it reaching its intended audience? Are the music rotations correct? How is the music presented? Is there a consistent personality and purpose, or is the station merely a jukebox playing songs in random rotation?
Then there is marketing. What was the marketing strategy? Where was the station promoted? How were potential listeners given a reason to sample it? Was the audio quality competitive?
Too many broadcasters approach AI primarily as a way to save money. They eliminate people, automate the presentation, put the product on the air and wait to see what happens. That is the wrong approach.
The first question should be: What is the magnet? What is the powerful reason someone would choose this AI station over satellite radio, AM, FM, another streaming station, a podcast or a personal playlist?
I have managed and programmed radio stations across the country—from the Northeast and South to the West Coast. Every region has cultural, musical and audience differences. If a station does not recognize and reflect those distinctions, it is missing the boat—regardless of whether its programming is created by people or artificial intelligence.
AI may be capable of keeping a station operating around the clock. That does not mean it can make an ordinary station compelling.
There is an old saying: garbage in, garbage out.
Let that marinate.
—Lee Michaels, The Radio Coach
Sources: Andon Labs’ live AI-radio experiment and RAIN News’ September 17 report.
What do you think? Can AI create compelling radio programming, or does successful radio still require human judgment, local knowledge and emotional connection? Share your perspective in the comments.