A recent TV programme posed the now-familiar question: Will AI take my job? As a musician and teacher, I watched with interest. After all, if an algorithm can teach scales, demonstrate pieces, and respond instantly to questions, where does that leave the humble piano lesson?
But perhaps that’s the wrong way to look at it. The real question isn’t whether technology can deliver information. Anyone with internet access can now find thousands of tutorials, slow-motion demonstrations, and “learn this in 10 minutes” videos. In many ways, this is great, and access to music has never been more open.
The illusion of progress
However, one of the biggest traps with this type of imitative learning is that it appears to yield results. You can follow along, copy what you see, and sometimes produce something that sounds convincing fairly quickly. But copying is not the same as understanding.
Without guidance, it’s very easy to reinforce mistakes and poor technique without realising. Studies and teaching experience alike show that self-taught musicians often develop bad habits, poor posture, or gaps in understanding that later limit their progress. There are obviously some notable exceptions to this, but these are few and far between.
A teacher’s role is not just to show you what to play, but to notice how you’re playing it, and to intervene before bad habits are adopted. I’ve spent a lot of time helping students unlearn bad habits before developing sustainable foundations for becoming accomplished musicians.
The value of a human ear
A video cannot hear you, nor can it notice that your wrist is collapsing, that your rhythm is inconsistent, or that your phrasing doesn’t match the style. It certainly can’t adapt in real time to your specific strengths and weaknesses. A teacher, on the other hand, listens, responds and adjusts. That immediate, personalised feedback is one of the key reasons students progress more efficiently with guidance.
Learning music, not just pieces
Another subtle limitation of copying online is that it often focuses on a particular piece or arrangement) rather than developing skills. Learning properly with a teacher develops a much broader skill set:
- Reading music fluently
- Understanding rhythm and pulse
- Developing technique that works across repertoire
- Listening critically
- Building musical memory and problem-solving skills
These are transferable skills. They eventually allow you to approach any piece of music, not just the one in front of you.
Structure, direction, and momentum
One of the least glamorous, but most important benefits of lessons is structure. When learning alone, you are the curriculum designer, the assessor, and the motivator. That’s a lot to juggle.
A teacher provides a clear path, set appropriate challenges, and keep you moving forward. Having someone expect you to play next week is a powerful motivator.
Independence, not dependence
A good teacher is not there to be needed forever. Their aim is to equip you with the tools to become musically self-sufficient, to diagnose your own problems, to practise effectively, and to learn new music independently.
Ironically, students who rely solely on copying often remain dependent on demonstrations and struggle when faced with something unfamiliar. True independence comes from understanding.
So where does AI fit in?
AI, like YouTube, is a tool. A powerful one. It can demonstrate, explain, and even simulate feedback to a degree. Used well, it can complement learning brilliantly.
But it cannot (yet) replace the nuanced, responsive, human process of teaching. It cannot fully replicate the judgement, experience, and adaptability of a teacher working with a real student in real time.
And even if it could, we might still find that what we value most in learning music isn’t just efficiency, but connection: the shared experience of making music with another person.
A vested interest!
Of course I do have a vested interest in saying that learning with a teacher is worthwhile. But if my job can be replaced by a video, then I’m not doing it properly.
Good teaching is not just about delivering information, it’s about developing musicians. And that’s something far harder to automate.