ART-ifical Intelligence

Over the last few years, there’s been much hype over Artificial Intelligence (AI) in art applications (Dal-E) and, more recently, literary production (ChatGPT).

As an artist who employs a fair bit of old-school appropriation in their practice, I’m open to the re-mix and repeat possibilities that AI offers to creatives…but I’m also wary as hell 😉

AI applications can generate images, music, and even stories, with the potential to revolutionize the creative process. However, as with any new technology, there are both pros and cons to using AI in the production of artwork. There are SO many existing articles that discuss the issues. I’m not going to re-hash them here, as I want to share something that I find much more juicy.

One of my accountability partners has been using Chat GPT to pen an elevator pitch for herself (hello Lisa! Psst, she also recommended the John Oliver episode I’ve linked to above). Lisa recently shared some of the results on her artist network, Make Big Art. After adjusting the prompts a few times, she found it delivered a paragraph that was a bit long, but that had great potential with some eloquent turn of phrase.

My curiosity was piqued. I created a free ChatGPT user account so I could explore the technology myself. I asked the program to “write a ~500 word article written in a casual tone about new york artists that use dioramas in their practice”. Faster than you can sing Bippitee Boppitee Boo, it spat out a competently written article.

Unfortunately, I wasn’t mentioned at all (oh, my bruised ego!). I tried a revised prompt, “write a ~500 article in a casual tone about new york artists that use dioramas in their practice that includes Canadian born artist Jody MacDonald”.

Because I asked Chat GPT to generate based on something I’m exceptionally familiar with (my own practice), I discovered a limitation that I found to be equally disturbing and delightful.

Sure, the article has pretty phrasing and some uncanny descriptive accuracy about my work. But there is also some downright, out-of-left-field, crazy confabulation (if you’re curious, you can read the full generated article below).

Having discovered this, I was interested to see how accurate the rest of the article was. I researched the other artists mentioned and then compared the information that the AI generated article created with what my Google search presented to me. After finding evidence that only some of the artists mentioned in the article work in the diorama genre, I realized that ChatGPT was making up all kinds of fake-news. Yes, that’s right. ChatGPT will make sh!t up and voice it with confidence and authority.

Now, I’m the first to admit that the inaccuracies about MY practice are likely attributed to my utter lack of SEO implementation (ChatGPT can only remix information that it can “find” on the interwebs). Maybe it’s possible that the other artists the app chose to discuss in its article are equally as SEO-inept as me. Or maybe this tendency to manufacture confident statements of pure bullshit is just an inherent flaw of literary AI.

Far from being upset at this quirk, I find it utterly fascinating. Of course, I would – I love all things ambiguous. The sliding scale of truth v. fiction finds its way into almost every piece I make, be it a work-on-paper, sculpture, or installation.

My experience with the software has me considering how I might use this technology within my practice – and not just in the obvious way (generating the text that I use in any miniature newspapers or magazines that might make up details of future dioramas). I’ve been thinking of using prompts to create fictitious articles or reviews about my work, and then create the work as it is described in these ChatGPT generated articles.

This is not a new idea for me.

True story: in 2007 my piece, Favourite Ways With Pheasant (Monday through Sunday) – consisting of seven textile figures — was described in a community tabloid as “rubber chickens dressed in lingerie”. My artist statement held no reference to rubber chickens; someone had looked at the image and assumed by their shape and position that the figures were rubber chickens.

I remember thinking how wonderful this bogus statement was and how fun it might be to start a series of works that were based on inaccurate descriptions of my art.

For a series like this to work, though, it requires that the artist’s work is consistently being both viewed incorrectly AND being written about incorrectly. For better or worse, the limited articles and mentions I’ve had in the press have been fairly accurate. Using ChatGPT, I now have the power to create as many wrong descriptions as I desire.

I can easily give prompts that will deliver fictional descriptions of art that are based on my existing, and maybe even yet-to-exist, works. I will then take these descriptions and interpret them into physical artworks. Potential exhibition title: How Many Lies Does It Take To Make A Truth?

P.S. In the three weeks I’ve taken to write this article, there has been a later release of ChatGPT. I have yet to try it, but apparently those who have explored it find it very good. I have to wonder, does that mean its false-truths are extra fantastical?

I sure do hope so 😉

P.P.S. Here’s the ChatGPT generated article from the prompt, “write a ~500 article in a casual tone about new york artists that use dioramas in their practice that includes Canadian born artist Jody MacDonald”:

Dioramas have been used for centuries as a way to create miniature scenes, often used for educational or entertainment purposes. In recent years, however, artists have taken this form to new heights, using dioramas as a means of expression and storytelling. New York is home to a vibrant art scene, with many talented artists pushing the boundaries of what dioramas can do. One such artist is Jody MacDonald, a Canadian-born artist who has made a name for herself in the New York art world.

MacDonald’s work is characterized by its intricate detail and masterful use of light and shadow. Her dioramas are often inspired by the natural world, with scenes of animals, landscapes, and even underwater environments. What sets MacDonald’s work apart, however, is the sense of movement and life that she is able to create within her miniature worlds.

One of MacDonald’s most striking pieces is a diorama titled “Arctic Meltdown.” The piece depicts a polar bear standing on a rapidly melting ice floe, surrounded by dark, choppy water. The bear’s fur is rendered in minute detail, with each individual hair carefully placed to create a sense of texture and depth. The water is equally impressive, with MacDonald using a combination of clear resin and blue pigment to create the illusion of movement and depth.

Another of MacDonald’s pieces, titled “Underwater Dream,” takes viewers on a journey beneath the waves. The diorama features a school of brightly-colored fish swimming through a coral reef, with shafts of light filtering down from above. Again, MacDonald’s attention to detail is remarkable, with each fish carefully sculpted and painted to create a sense of individuality and life.

MacDonald is not the only artist in New York using dioramas in their practice, however. Many others are also exploring the possibilities of this versatile art form. For example, Anne-Katrin Spiess creates dioramas that are inspired by old-fashioned toy theaters, with whimsical scenes featuring dolls, animals, and other small figures. Meanwhile, Lori Nix and Kathleen Gerber create dioramas that are more dystopian in nature, with scenes of abandoned cities and ruined landscapes.

Overall, the use of dioramas in contemporary art is a fascinating trend that is worth keeping an eye on. In the hands of talented artists like Jody MacDonald, dioramas can be used to create incredibly intricate and immersive worlds that captivate the imagination. If you’re in New York, be sure to check out the work of MacDonald and other diorama artists – you won’t be disappointed!