top of page
david003464

How AI can help improve radio news bulletins



In the 2000s I often did shifts with Smooth Radio in London, covering news for the capital but also recording news bulletins for other regions in the network. This meant recording two, sometimes four, bulletins an hour, plus doing two live ones. As most newsreaders know, you tend to make more mistakes when recording bulletins than when it is live. There is something in being live that focuses the mind. So every hour would feel like a treadmill of writing, recording and editing, writing, recording and editing, writing, broadcasting. And of course, there was no time for a break in a nine or ten hour shift.


As a consequence the time you invested in those regional (recorded) bulletins was minimal. You would inherit some scripts from a colleague in the regional team and just record what they sent you, around other latest stories. There was also no time to research the news for London properly and craft good/better scripts. You often used IRN (the industry newswire) which was not always written well. This type of work pattern is very common on radio shifts at the larger networks or stations. These days there is also online content to do and podcast material to consider as well. The pressure on a newsreader shift is constant.


In the modern newsroom this type of churn could be rectified by embracing AI. There are already companies providing news that is voiced using the technology and I only see this becoming more common. But let me be clear at this stage, I completely automated news bulletin with an AI voice is not what I recommend. As someone who has read the news on radio and in podcasts for 25 years, I want to protect my 'role' of course. So what I suggest is newsreaders being given the opportunity to focus on the important things and as a result produce better content. Remember the adage that AI will not for certain replace you, but someone who understands AI better probably will.


Having said that, I do hear a lot of radio bulletins that would be better voiced using AI. Newsreading is an art and requires proper training. Sadly the loss of smaller radio stations and a change in university courses means that the development of skills happens much less than it should. Masters students will typically get little more than an hour of voice coaching during their course and graduates are being thrust into national roles before they have developed their voice properly because there are no stations to learn at. But the problem is also in the writing of scripts. A great radio bulletin, requires a great script and that requires time to craft it.


Imagine a role where a human is overseeing the newsdesk and allowed the time to write and work on the flow of the bulletin. Be more creative perhaps. To be honest, radio bulletins are very old fashioned. They have not changed in 60 years. Instead you could get AI to have your voice 'banked' and ready to 'read' any script you feed into the system in your style. This way you could be writing, editing and crafting news bulletins for the whole hour and drop a script into the AI system seconds before it goes out. Effectively it would be 'live'. The other major plus would be no stumbles or fumbles. No pronunciation of words wrongly or even place names if the system is updated properly.


In terms of workflow this would make the newsreader role a lot less stressful and allow them to focus on the strengths of having a human on the desk. Fact checking, editorial decision making and crafting the structure of the bulletin. There may even be a way to use the technology on the move, so that you could cover a breaking story and still deliver these bulletins from the field.


There will be concerns about this leading to fewer jobs, but what I describe is not about replacing someone, it is about helping that person do their job better. After all, the voice is precious and legal protections would be in place to make sure your voice was not used without permission and only for certain content.


What I advocate is you still owning that voice and improving the way you do your job. Some bulletins would still be live perhaps and you could be reporting from events as they unfold, because AI helps you to deliver bulletins. Currently many newsreaders are just churning out content with very little time to do proper journalism, like those shifts I did for Smooth in London. This could mean humans being deployed to do that proper journalism I describe. Get AI to do the churn and allow the human to be the journalist - and do the journ!

Comentarios


bottom of page