Google has done it.  They have created an AI that can generate music from text descriptions.  The announcement of this is timely because recently Chat GPT, (a language learning model created by Open AI), can do everything from writing school papers, song lyrics, to solving coding problems, and is dominating the news with reports that this tool will be a massive disruption for school professors, marketing companies, programmers and creatives alike.

‍

Google’s learning model is called MusicLM.  It has the ability to produce songs that are particularly complex in composition and fidelity.  It was trained on a dataset of 280,000 hours of music to learn to generate songs from simple and complex text descriptions.

‍

Attempts at this AI have been tried before, however with MusicLM there is a major difference.  The songs sound good.  It is reported that the music produced is able to capture human nuances like instrumental riffs, melodies and moods.  The system can also build on existing partial melodies that need to be resolved, and can pick up on this by receiving inputs through humming, singing or whistling.

‍

The Pros:

‍

One of the key advantages of this AI system is its accessibility. People who have no musical background can now create music simply by inputting text. The AI system takes care of the musical arrangements and generates original compositions that can be further refined and modified by the user.

‍

The technology also offers a significant time-saving benefit for experienced composers. The AI system can generate musical ideas in seconds, freeing up more time for the composer to focus on refining the pieces and adding their own personal touch.

‍

The AI system is versatile and can generate music in a variety of styles, from classical to pop and everything in between. This means that it can be used for different purposes, from creating background music for videos to composing original soundtracks for movies or games.

‍

‍

The Cons:

‍

It is important to note that currently, the MusicLM system is not without limitations. The music generated by the system is not always emotionally engaging or compelling. If it is used by a composer, it will be important for them to add their own creativity and emotion to the pieces generated by the AI system.

‍

Another thing to consider are the difficult questions surrounding the legal and commercial implications:

‍

What constitutes an original composition, when the music created is either AI assisted or purely made from AI?

‍

After learning from human data sets, could the AI create melodies so similar to existing music created by humans, that it crosses the line of our copyright laws?

‍

Will the average listener care whether the songs they listen to are created by a human or computer? 

‍

What does this all mean for the millions of music creators around the world creating to make a living?

‍

These are all questions that can be answered over time, but for now, because of the ethical challenges and risks, Google has decided not to release MusicLM to the public.

‍

This is a relief for the present day music creator, but in due time MusicLM or something like it will launch and it will continue to improve in performance. One can only imagine what the world will look like when the machines are able to create music so well that the job of the composer, producer, songwriter and artist becomes unnecessary.

‍

‍