In this episode, I'm going to be exploring the use of AI-generated music in the classroom. The AIcademia podcast is a weekly show helping educators like you leverage AI in your everyday practice. I'm your host, Andy Fisher, and thanks for joining me.

Over the past couple of years, AI-generated music has come on in leaps and bounds. Platforms like Suno and Udio allow users to create original tracks - either instrumentals or songs with lyrics, in a matter of minutes, making them highly accessible for educators looking to integrate music into their lessons. 

If you're new to this, here’s a quick primer:
Both Suno and Udio offer a limited number of daily song generations even on the free plan. For example, Suno provides 10 credits per day, and each time you generate a song, you receive two different versions—so you can create up to five songs daily. If you then choose to invest in a paid plan, I have found it excellent value for money and I can generate more than 500 songs a month for just 10 dollars which means that I can experiment without being concerned about running out of credits. The platforms work using built-in large language models that generate lyrics, which can then be set to a variety of musical styles. Alternatively you can upload your own lyrics if you have something specific in mind.

If you have a musical ear you can fine tune your prompts to create a particular genre of music or you can come close to duplicating one of your favourite bands. I was pleased to see that if you name a particular artist the model will refuse to execute the prompt and will give you a message that explains it would be in breach of the community guidelines. We'll come back to the ethics of AI music later in the episode. But if you want something in the style of say Ed Sheerhan, you might prompt for an acoustic guitar track with a male British singer songwriter and you’ll get something a little like this: [song excerpt]

There are some handy suggestions on the platform on how to create unique musical styles or you can just hit the 'roll the dice' option to randomise the output. So you might go for a modern country and western track with male vocals or a 90s glam rock anthem, you could try for an epic cinematic fantasy theme or a musical theatre duet and each of these prompts will create very different results.

It does sometimes take a little experimentation to land on a song that matches the ear worm in your imagination but that's part of the fun for me. 

For more advanced users, you can try meta prompting by adding instructions in brackets like 'verse', 'chorus', 'guitar instrumental' or 'bridge' - the models don't always follow these instructions to the letter but it does give you a little more control over the shape of the song generated.

Once you've submitted your prompt, the process is really quick—songs are generated within 30 to 60 seconds and can be listened to in the app.
The ease of use and quick turnaround of AI music generators like Suno make them practical not only for creating resources in advance but also for real-time use during lessons.

so let's explore some examples of how I, and my colleagues have been using AI music generation in our lessons this year so far.

My wife, who teaches primary is currently working with 7 and 8 year old pupils, has used AI-generated songs to help with times table memorization. Here is a clip from a track we made in Suno - the lyrics and music being created on the platform: [song excerpt]

She also produced a hip hop track to introduce a new scheme of work on the various regions of Britain. This was used as a backing track as the pupils walked around the classroom scanning QR codes with their tablets. Those codes took them to colourful landing pages with facts and deatils about those counties - and yes, we created those websites using an AI no-code tool called Replit which I can talk about in a future episode. Here's a snippet from that song: [song excerpt]

As a secondary school English teacher, I’ve used AI music twice this term. In the first instance, I used it live in a lesson exploring the presentation of the witches in 'Macbeth'. Having read the scene in whicb the wyrd sisters cast a spell around a cauldron, pupils were tasked to work in small groups to direct the scene for a modern interpretation of the play. This included setting the spell to music.

We had some great ideas. One group suggested the witches should appear as ageing rock stars like the Rolling Stones, another group wanted them to be Goth K-pop girls and a third group suggested a headbanging death metal vibe.
I fed each of these styles into Suno along with Shakespeare's verse and then played the generated tracks and we discussed how successful each version was and whether they would work on the stage. Here's an excerpt from one of the outputs: [song excerpt]

The more recent time I've used AI music this term was to support my son in completing a homework about yeast. He had to make a fact sheet and wasn't feeling particularly inspired so we sat together and used Chat GPT to create some suggested lyrics. We then tweaked them until we were happy before uploading them to Suno and here is the result: [song excerpt]

We forwarded a copy of the song to his Biology teacher who was so taken with it that he played it for the whole class and is now using AI music as part of his classroom practice.

I can imagine that other subject teachers will quickly find use cases that suit their own classroom content. Modern Foreign Language teachers might use this technology to introduce new vocabulary. Drama & Theatre teachers could create original and royalty free background tracks for their class productions and since we often remember song lyrics more easily than raw text, AI-generated songs could be used to reinforce key concepts across the curriculum as part of a revision process.


Finally I'd like to mention a half term's scheme of work using AI Music Generators that I've just launched with my Year 10 students who are part of the pilot 'Introduction to AI Course' I'm running which is a non-examined alternative to GCSEs. I mentioned a little about this course in episode 5 last week. 

In the next few weeks I will introduce Udio, Suno and some of the other AI music platforms available. Students will learn how to use them and will then work towards generating an original song on a topic and in a style of their choice. These tracks will then be submitted anonymously to a member of my School's Music department along with a few lesser known human generated songs. 

My brave and open minded colleague has then agreed to see if he can figure out which are AI generated and choose the track that he feels is the most successful piece of music submitted. The project will culminate with all of us coming together to discuss the experience and broadening the conversation to consider the ethics and implications of AI music.

We'll be asking thorny questions like these:


What are the data sets from which AI platforms like Suno and Udio draw to build their algorithms? Have human artists had their materials scraped and used without consent?

Are AI music generators a good thing because they democratise the process, allowing creators to produce and share their work with a global audience without having to deal with the gatekeepers and record labels that have traditionally controlled the industry

Or are these platforms reducing the opportunities for musicians to earn a living as the market becomes flooded with cheaply produced tracks made by users who have never played a musical instrument or spent a minute in a recording studio? 

And are AI generated songs original creations or just the musical equivalent or spam - reconstituted and homogenised output that on balance dilutes human creativity?

Some very thought provoking questions which I’m looking forward to wrestling with.

So, once we have completed the half term's work, I'll produce a follow up episode to let you know how it went and perhaps I'll even be able to persuade my colleague in the music department to join me and record a conversation with these topics in mind.


To round things up for today, and putting aside the broader implications of AI music to the music industry itself, I think platforms like Suno or Udio are exciting, creative tools that can bring a fresh dimension to teaching our subjects. They're free, easy to use, and I'd encourage you to experiment with them yourself.

Thanks for listening—I hope you’ve found some useful takeaways from this episode. Please spread the word if you think others would enjoy the show, and don’t forget to check out the AIcademia YouTube channel for practical tutorials on using AI tools in education. Have a great week, and I look forward to catching up again soon!