You could have done this with AI

Various tests of how AI performs over humans have found errors with how artificial intelligence completes various tasks. We are quick to point out flaws in the technology, sometimes because its fun to look at seemingly obvious misspellings or visual flaws, rather than ask the hard questions about what this boom means for our world and our humanity. The fact of the matter is, while AI systems were initially trained on administrative and technological tasks (such as conversing over slack or assisting with customers on corporate platforms) rather than creative ones, Large Language Models (LLMs) are learning fast. They are competent at medical diagnoses, human conversation, and are now considered to be even more creative than humans.

You’ve heard it before and you will hear it again. The question laced with concern and anxiety about whether or not AI is going to take over this job, that project, or yet another billboard near you. With the rise of the panic of AI takeover, professionals are rushing to explain that, no, AI can’t do what they do because the process of their work is uniquely personal, or requires very specific understanding, forethought or design. It contains that unique human element that people struggle to precisely articulate but is uniquely them.

While I am happy to think and write about artificial intelligence across fields and systems, my procrastination (writing this) can only last so long. So I will instead focus on how AI has been encroaching on one of my favorite hobbies and pastimes: art, as well as what I think that means for the intersection of my passion (art) and my job (science).

Art, Science, and AI

Last week, the journal Nature focused on an often overlooked aspect of science: illustration. Science illustrators provide creative visualizations of things too big or too small for us to grasp, work closely with researchers to develop images of time long past, and otherwise serve as essential communicators. A common phrase heard in art classes goes “draw what you see, not what you think you see”.

Scientists may be familiar with this core concept: hypothesis driven research is designed to frame a specific question and objectively report on findings: a hypothesis serves as the reminder what what they “think they see” while results tell us what was truly there. The world of art, like science, allows us to better understand the world. It was, after all, the famous artist-scientist Leonardo da Vinci who is thought to have grappled with theories of acceleration and gravity long before Galileo.

The form of attentive seeing practiced by artists and scientists alike is what shows people new perspectives of the world as we know it. This is an incredible feat on its own.

The interview-style article in Nature was full of well-worn justifications defending illustrators’ role as professionals: their work takes time, expertise, attention to detail. Just as their colleagues in other fields have attempted to do, illustrators populated their recent interview with increasingly specific abstractions of why AI could not and would not replace them. The reality is, if AI can’t do these things yet, that is not a guarantee it wont be able to at some point in the (likely very near) future.

So, why do it?

In the name of class time and resources, the growth of STEM has wrenched science and art apart. But if Leonardo da Vinci has taught us anything, its that the separation was never destined to last. The increasing movement from “STEM” to “STEAM” is a modern testament to that, and begs the question of why the humanities and science were sequestered in the first place.

For me, drawing serves at a meditation. Before the uncertainty of our AI-dominated future, time in an art studio, or on the living room floor with paper was never wasted. At the end of the hour or two (or six), something new would exist in the world. For the most part, I was not very good. Images wouldn’t turn out quite as I pictured them, would take laborious hours to construct, or change altogether to accommodate the fact I was not able to execute my plan A. Which brings me back to the elephant in the room: if at some point, an LLM may be able to do it better than I would, why keep doing it?

Firstly, AI is not there yet, and the mistakes made in images used for science communication can be misleading and dangerous to science education. But, to me, the most important piece of the puzzle is the passion, skill and interest of those still engaging with the art form. I think matters for people to hone their tactical skills and further this engaged form of seeing that scientists also pursue in the hopes of deeper understanding. This pushes us to be better scientists, better artists, and more engaged humans. Plus, its fun - shouldn’t that be enough to keep carving time for it?

The article was coupled with another, by the same author, focusing on the joy of their work. It is in this joy that I want to focus. When it comes to creative acts such as the role of AI in writing, or illustration, it feels that we are asking (and justifying our frail answers to) the wrong question. It’s not “will AI replace me?” I think the question should instead be, “why am I doing this?” And if that answer is simply: “I enjoy it”, that feels like enough justification for me.

What next?

I am currently feeling inspired and encouraged by the rise of art integration to science, two fields I feel passionately about and have always felt are deeply connected. But it doesn’t have to be art. In the era of AI, it will become increasingly important that time is spent pursuing passion, skill, and things that make the world feel a little more human.

To point out the elephant in the room - I could have told a LLM to write this for me. It would have done so faster, more clearly, and with fewer grammatical errors and fumbling sentence structure, I’m sure. But that is not the point. I enjoy writing and thinking about topics that interest me. Why would I outsource that work to a computer when there is so much I successfully procrastinated in the time it took me to write this? (to my brother, if you’re reading this, I promise the officiant address I gave at your wedding was not one of those procrastinated tasks…)

If you’ve read this far, congratulations, you are fighting the tide of AI summary-induced attention spans - I’ve been thinking about this topic a lot and would love to talk about it, feel free to reach out!

Looking for more art-science resources? Check out my portfolio page for more links!

Need art for your science presentations or papers? I would love to help so please reach out!