Coming up in two weeks is a free and open microcourse in which my co-teacher and I will explore making images using A.I. specifically for instruction. Thanks in advance to BCCampus for hosting this course.

REGISTER HERE: FLO MicroCourse: Artificial Intelligence Images in the Classroom.

I’ve had a chance to learn more about A.I. image generation while researching for this course, and as an artist, I have to say my initial skepticism about these tools has not been proven wrong. At least in terms of the quality of the final output of the images in question. They are all just too ‘artificial-looking,’ if you know what I mean. However, I am finding the process for creating these images FASCINATING. I believe this has to do with the fact that it involves using words and language to create the images as opposed to your eyes and body. I am approaching this subject as an artist myself, and with some reluctance, I can say there are very few artists or image makers that will produce visuals simply from producing a text prompt. At least from a fine art perspective, it involves using all your senses to observe the world around us and then recreate a ‘vision’ in some media. There is a back-and-forth. You respond to your marks as you are making them. It is a journey.

Not with A.I. In this world, the text prompt rules all, and it is here that I find some potential for using A.I. as part of a creative process. I mean, I have studied and practiced art all my life. I know the history, I know the culture. I have practiced and participated in many creative activities and learned the language of art-making. It seems to me that untrained folks who want to use A.I. tools to make images, the tool will have to understand some of the basics, at least if they want it to work effectively. But look at that old guy who is talking.

For our micro-course, we are not focussing on the ‘how-to’s, nor are we doing a software comparison to make any recommendations. Amongst our learning objectives is that our participants “Develop a personal framework for ethical decision-making with the use of AI in artistic practice.” Something I am eagerly putting together myself.

To that end, I have begun some of my own experimentations and out of the gate had a pretty tough challenge for an artist or A.I. to work with as visual prompt. Here are the results of my first challenge. I am sharing the prompt and 5 of the results from different A.I. tools, mostly the free ones. Dall-e. Firefly. Co-pilot. Deep Dream Generator. Leonardo.ai.

Prompt: I want to create an image that shows the tension between a traditional artist and artificial intelligence. This should look like an everyday photograph that somebody could take with a regular phone. I would like it to use a visual metaphor for this tension, not using a traditional-looking artist or any people in the photo at all.

I used the exact same prompt in each platform and did no further modification to the image.

Clearly, there needs to be more back-and-forth with me and prompts to make the output something I would consider ‘creative’. In many cases, the A.I. could not interpret my meaning correctly, like when I asked for no people in the photo. But now we are getting down to semantics, aren’t we? How often do I want to redesign this prompt to get a result with no people in it and not use traditional artist tropes like paint brushes and easels? Can I ask the image to develop a metaphor independently, or will I need to provide that myself?

I guess the point I am finding my way to, is as a visual artist, how much time do I want to spend describing in words the image I want to see with my eyes as opposed to sketching ideas, reading through inspirational references and exploring the world outside to come up with a striking image to show “the tension between a traditional artist and artificial intelligence.