Even music, one of the most innately human things, is able to be generated by AI now. SoundDraw is an AI music generator that can be used to create music. It is limited to instrumental music for now, there are no options for generating vocals. There are many customizable options like genre, mood, tempo, length, and instruments. Using these options, a variety of tracks is created. After generating you can still make some changes to the segments.
In my experience the quality of the songs varied quite a lot. Some songs were interesting, most of them quite bland, and quite a few of the songs sound off or weird. The AI does manage to get the basics of the genre correct most of the time, but there have been some hiccups here and there like rock songs with just synthesizers and techno without any bass. The keywords that are used to describe the mood were mostly just plain incorrect in my experience. If you asked me to describe the mood of a song, I would often have given a different answer than the attached keywords.
This tool could be very useful for background music in videos and advertisements, but the quality does not come anywhere near human musicians or producers. The tool is marketed and licensed as background music in their subscription plans. I think tools like this could compete with the libraries of royalty free background music that are the standard for background music at the moment.
Since generative AI does not actually generate any truly new ideas, you might be wondering about the originality of the songs. The songs might start too similar to existing artists which could lead to licensing issues. SoundDraw has addressed this issue by training their AI on their own datasets that have been made by their own producers.
Ultimately, I think that AI generated music will not go much further than background music in the near future. I encourage you to try the tool for yourself and see for yourself what it has to offer.
Very cool AI tool, and I like that you reflect a lot on your own process and opinion of the music. Looking at the progress of music, it does feel like AI is the next logical step: from acoustic, to electric, to synthesizers, to software and now to AI. It took a while for each step to get adopted, so maybe AI generated music in the future will be something as popular as EDM-producers are nowadays. Personally, I think that humans transcend technology with their creativity and emotions and I hope AI will never be able to replicate that. It’s nice to hear, at least from your perspective, we’re a long ways from that future!