On the Use of AI for Creative Arts

AI is being pushed on us from every direction with prophets proclaiming it as either the savior or the destroyer of this world and our way of life. Here I want to offer a reflection on the use of AI for creativity. This article is being written (at least partially) as a response to Paul Kingsnorth’s article Writers Against AI.

Many thinkers I respect – including Jonathan Pageau and Paul Kingsnorth – see AI as extremely dangerous. I agree with them on several points: it very likely can be used as a conduit for non-embodied consciousness (i.e., spirits) to interact with us, and it is sadly being used to replace human creativity and thinking. To the first point, I have no interest in receiving spiritual direction or advice from an AI bot. To the second point, I have little interest in creative content generated by AI. While I’ve used it to generate a few images on this website, I do so only because of copyright laws restricting the usage of human-produced art. I always prefer art made by humans.

I read or heard somewhere recently that AI is averaging everything to mediocrity. Blog posts, news articles, videos, stories, etc. are being produced by AI. Because a machine can’t possibly be creative (it can only copy ideas that it steals from humans), it raises up humans who have less talent and drags downs lazy but creative humans to a widespread mediocrity.

Obviously, this is bad for human flourishing and creativity. We humans were made in the image of God, and part of that image is being creative as God is the source of all creativity. I think in response, we’re not going to see all people mindlessly sucked into using AI for everything. Instead, we’re going to see a fairly large contingent of people around the world who stick to websites, books, and publishing platforms that prioritize human sweat and effort.

I’ve used AI quite a bit, largely for bringing together multiple sources so I can research a topic. Other times, I use it for troubleshooting technical issues in Linux. The results overall are admittedly mixed. Sometimes it swiftly finds a solution to a tech issue; other times it creates more problems than it solves by offering bad solutions. The research findings that it presents are sometimes sufficient, and other times lacking in important nuances.

I like Paul Kingsnorth’s manifesto Writers Against AI. But at the same time, I’ve chosen not to subscribe to it fully or put the logo on my website. Why? Because I sometimes use AI to proofread my essays and articles. I’m a busy father of young children and a full-time priest. My training is in theology and business administration, not in writing. While I produce my own content and often sit on it for a week or two, proofreading and revising it, I appreciate having a machine do a final proofread to check for stupid mistakes, grammatical errors, and other oversights. I’ve always enjoyed writing, but at this point in my life, I don’t have time to undertake a formal education in creative writing.

While I fully support the Writers Against AI idea, I find it too limiting for myself and many other people. It’s like saying we should begin a Writers Against Technological Assistance group that doesn’t allow spellcheck, grammar check, or even computers. Everything must be written by hand or with a typewriter. Such a goal would be fine for those who wish to embark on it, but for most of us, it’s simply too restrictive. Admittedly, some of my best homilies (sermons) have come from me sitting down for hours with a pencil, notepad, and a stack of books. There’s a creative element we lose when we utilize computers for writing. The same could be said for the assistance of AI for even simple things like proofreading – especially if the author either lazily applies all suggestions or doesn’t take the time to learn from his or her mistakes.

I remember decades ago when Antares released AutoTune (a digital vocal pitch corrector). Many people decried it as a tool for talentless singers with an attractive face to be successful. And it continues to be used exactly for that purpose. In those days, I wrote and recorded music frequently. So, I decided to give this new tool a try. Sometimes I used it for the wacky effect it would give my voice, and sometimes I admittedly used it to fix a vocal track I didn’t feel like re-recording. But I often used it to hear where my pitch was a bit flat or sharp and I needed to re-record part of a vocal track. Would it have been better to go back to school and get a master’s degree in music theory or voice? Or perhaps spend years in intense vocal training with a professional teacher? Yes, but music was a hobby, not a profession for me. It was a creative outlet, a folk art and musical diary of sorts.

My point of this article is that while we should be cautious about not offloading our creativity to AI, I don’t see much use for the common person to be an ultra-purist about it. Those who are on the high end of the talent spectrum (like Kingsnorth and Pageau) don’t need it. If anything, it would get in their way. But for those of us who are not highly trained and are not seeking to make a living from our art, I don’t see the harm in a careful, selective use of it.

It’s hard to make a manifesto from my position, though. I’m not taking a hard line in either direction. Who would want to subscribe to Writers Against AI Over-Usage…But a Little Is OK? Maybe I’m wrong about all this. It’s a new technology and we’re all still trying to figure things out. I’m certainly open to correction. And maybe in the coming years – if AI annihilates human creativity – I’ll agree that Kingsnorth’s purist approach was better.

End Notes:

Featured image courtesy of Getty Images. Ancient illuminated manuscript from about the year 1255 depicting the dragon giving the scepter of power to the beast from the sea. Since much of the discussion having to do with AI is apocalyptic in nature, I decided to roll with that.

Also, I used a private, self-hosted AI to proofread this article, asking it to maintain my voice and wording. It suggested I add italics in the various places that you see above, add a comma where I missed one, and add a hyphen to “over-usage.” To me, these kinds of changes are minimal and a good use of AI. It’s not producing any content nor is it offloading my creative process to a machine. It’s not much different from using Microsoft Word’s spelling and grammar check, which people have been using for many years without controversy.