Using AI for creative skill building
While chatting a few months ago, a friend of mine who works in the tech industry expressed a clear impression that AI will, without much doubt, be a prevalent tool in the future.
I argued, “when I use AI, all it gives me is anxiety about inaccurate data.” My friend responded, “well, that’s just like Wikipedia. It’s not supposed to be taken at face value, but we still use it to get a start on research anyway. It’s meant to make things more efficient” (not an exact quote, but it summarizes what he said pretty accurately).
It’s no secret that there is some heated debate among creatives regarding involving AI in artistic works. However, after this conversation with my friend, I decided to reapproach using AI with a different mindset. How can AI help me as a creative?
AI will probably be like to kids right now what browsing the internet or exploring a new app was like for me and my friends when we were younger. I remembered how I explored software tools as a kid and decided to try approaching AI tools that way.
Recently, I’ve been hungering to create a TV show OST (original soundtrack). After playing piano for some years now, I fell in love with the idea of creating music to an original scene. To fulfill this new bucket list item, I have been working on developing a whole original film to produce (more on that later).
The daunting thing about trying something like this, however, is that I had never done it before. My plan was to test this skill when making the film. However, what if during the post-production process, after all the effort’s been made, I realize that I’m not ready? It was starting to feel like a high-risk plan.
When I was a kid, practicing my skills didn’t require much risk or expense. Practicing magic only took a deck of cards and some YouTube tutorials. That was enough to get myself ready for some street and stage performances.
While thinking about AI a few weeks back, I thought, if all I need is some original video footage, maybe I can use AI to provide me some.
The biggest challenge was finding a suitable AI video generation tool online. Turns out most AI tools, particularly free ones, are made to generate very specific types of videos that most creators look for, such as TikTok videos with voice-over narration or explainer videos. In spirit of my childhood habits of learning and exploring new skills for free, I searched for around three hours for a suitable free service. Eventually, I caved and coughed up $20 to OpenAI’s Sora.
The goal was to let AI provide me a fresh video to make music off of. I wanted to treat it like training so that I could build a muscle of coming up with music based on something original. I inputed a brief prompt, and Sora spit out a 5-second video (per the Sora’s limit — I looped it on iMovie later for a longer duration for OST). A melody came into my head after watching it so I went to my piano and worked it out. In spirit of my first time doing this, I did not have a MIDI cord or a program to record the piece straight from my piano… so I recorded it on my phone’s voice memo (excuse the subtle squeaks here and there). You can check out the result with the video hyperlinked here.
I remember a quote I read on social media (I don’t remember exactly who) that went something like “I don’t want AI to do my art so that I can do my dishes; I want AI to do my dishes so that I can do my art.” Many artists are against AI creating art for them because art is a passion for most, not just a job. If AI is to be involved in an artist’s work or lifestyle, maybe it can contribute as a tool for training, experimentation, or education.
Though this experiment was quite rugged, as a first attempt, I don’t think it was all that bad. This exercise did allow me to take original video footage and practice creating an original soundtrack based on the footage. I do believe it was a good start. I will have to see how this purpose can be further refined.
