We used to run AI-generated jingles.

Nobody complained. They did the job. They said the station name, they sat neatly between shows, and they cost nothing but a few minutes at a keyboard. Then we recorded new ones with real people, and the difference was obvious straight away. The AI versions had been fine. The human ones sounded like us.

That small change is the whole argument, so I want to start there rather than pretend we have always been above the question. We tried it. We heard what we got back. We changed our minds.

What the industry is arguing about

AI in radio has moved quickly from a curiosity to a business plan. Stations can now generate voice links, weather updates and whole shows without anyone sitting behind a microphone. Some are doing it quietly.

The radio commentator James Cridland has been tracking this closely, and his position is one I agree with. Speaking at the FWD Conference in June, he argued that music is no longer the reason people choose a station. “It’s the people, it’s the DJs, it’s the hosts, it’s the shows,” he said. “Human connection, a shared experience, a club to join, a place to belong – that’s why people are listening to their favourite radio station now.”

If that is what listeners actually come for, then automating it away is a strange thing to do. You would be removing the one part of the product that streaming cannot copy.

Cridland has also written about what happens when stations get this wrong. The Australian station CADA ran a presenter called Thy for six months before the truth came out in 2025. Thy was a cloned voice, and the station never told anyone. The story only came out because someone noticed that a supposed 20-year-old radio host had no social media presence at all. Cridland’s verdict was blunt: “This breaks the number one rule of not lying to your audience.”

What strikes me about that story is not the technology. It is the assumption behind it. Somebody looked at four hours of daily airtime and decided that the person in it did not really matter. That the audience would not notice, and would not mind if they did.

Where our line sits

So here is our position, stated plainly, because I would rather be held to something specific than to a vague promise.

HCR will not use AI to make shows. Every voice you hear on air belongs to a person with a connection to Harrogate. We will not use AI-generated music, and we will not use AI-generated images.

We have used AI to help sort out problems while building the website, and I am not going to pretend otherwise. Fixing a broken bit of code is not the same as inventing a presenter. The line is the thing you hear and see as the station.

HCR is not a general purpose media company. It was built by three people who wanted a local platform, and its identity rests on being local and human. Those are not features we can automate and keep.

What that costs us

Staying human is not free, and a piece like this is worthless if it pretends otherwise.

We broadcast 24 hours a day with around 40 active volunteers, none of whom are paid. That maths does not always work. There are gaps in the schedule and there are repeats, and an AI presenter would fill both instantly, at any hour, without ever getting ill or busy or fed up. If our only goal were to keep sound coming out of the speakers, the case for automation would be a good one.

The audio is not always clean either. Our shows are recorded at home, in spare rooms and on kitchen tables. You will hear background noise. You will hear people lose their thread and pick it up again. We leave that in. Not because we cannot edit, but because the wobble is evidence that a person made this, in a real room, in this town.

Commercial radio spent decades sanding those edges off. AI can sand them off completely. What you get is smooth, consistent and nobody’s.

Why we think it is worth it

Our listeners tell us the same thing over and over: they value a local voice. Not a voice that sounds local. A voice that is local, belonging to someone who walked down Beulah Street this morning and has opinions about the use of The Stray.

You cannot generate that. You can imitate the accent and you can feed it the street names, but you cannot generate the fact of a person choosing to spend their Sunday afternoon making a two-hour show about free jazz for no money, because they want other people to hear it.

That is what we are protecting. Not a rule about technology, but the reason anyone tunes in.

Harrogate Community Radio broadcasts online and is available on TuneIn and Alexa. If you would like to get involved as a presenter, volunteer or supporter, get in touch at info@harrogatecommunityradio.online.