A Retrospective on AI and Creativity

Image Source: https://srinstitute.utoronto.ca/news/what-is-the-future-of-ai-alignment

When you hear “AI and Creative Arts,” what comes up for you? Is it excitement? Confusion? Frustration? A little bit of throw up? It’s okay to be honest. In fact, I would say today’s post is inviting you to be honest.

A lot of the same frustrating emotions came up for me when I was planning a recent event for the Rocky Mountain AI Interest Group (RMAIIG). Many can tell you: I grappled with the topic a lot leading up to the event. How do you make an event about something like this that actually serves creative individuals, when many feel like there is no intersection to begin with?

Today’s post is somewhat of a retro/introspective on the event, and what I learned in the process. Let’s dive in!

Wait, there was an event?

Yes! You can see the event details on Meetup, but essentially:

• About 170ish people registered, and at least it felt like that much in the room.

• An exhibition with just around 20 community members showcasing their creative works. The lab in the CU Boulder Aerospace building was jam packed, and we had prizes for 1st, 2nd, and 3rd.

• Two amazing speakers, Mark Johnson and Dr. Lee Frankel-Goldwater, sharing their experiences navigating AI as creatives. While Mark had embraced using Midjourney (an AI image tool) after a LOT of stress and ongoing turmoil, Lee took to defining what actually makes art…art through a process he calls the “Dickinson-Turing Test.” Both were phenomenal.

  • The quote for the article’s subtitle actually came from Lee bringing this up in his presentation, and I thought it was fitting.

We created a virtual gallery for every submission to this event here: https://rmaiig-art.lovable.app/.

Rather than talk at you about the event, I think it’s prudent to talk about my thought process and where I grew in the process.

AI and Art is not Black and White

When I came into the event, I had a hypothesis: this event was going to be a disaster. Not because I couldn’t trust myself with execution, but because AI and Art is such a divisive topic. I shared this in my intro during the event: while AI art as a search has increased by 1600%, the Project Management Institute (PMI) claims 70% to 80% of AI projects fail. Tools are being built, like Glaze, Nightshade, and Anubis, to circumvent model scraping or prevent bots from accessing websites, full-stop. Meanwhile, most creatives feel like they can’t ever survive making art as a profession. How do you invite people to contribute to an event like this? Frankly, a tall order in my mind to not get carried out with pitchforks and torches.

I did my best to find meaningful speakers, tailor the event details as I felt comfortable, and pushed forward. What was interesting when reaching out to countless individuals to contribute was that, despite my preconceptions, creatives were engaged. They were actually surprised I even wanted their opinions. And several of them submitted, despite the fact we were an ‘AI interest’ group. There were concerns about AI model training that I addressed, but the overall expression was excitement and curiosity. That got me thinking, how many people were actually just ignoring feedback instead of talking with creatives? Likely more than we realize.

The more I went on, the more I started questioning how technologists approach implementing AI. Had a couple of exhibitors come after the fact to let me know a certain someone I know was berating them on the ‘alarmism’ of their submissions. So much for free speech.

It’s not fair to write people off as Luddites when you yourself aren’t listening either. This ended up materializing into one notion: we’re not bringing everyone that is involved with AI to the table nearly as much as we should.

Human-first AI is where the magic happens.

In a similar vein, I was equally surprised as the submissions came in of the sheer quality. AI or otherwise. And that’s not me giving anybody a participation prize, either.

There were songs made with AI. Paintings criticizing and foretelling omens of what unfettered AI means for us. Letters translated from long-distant relatives using AI that would otherwise be significantly more difficult to connect with. Interactive Videos. Explorations of bias (ageism, mainly) in how AI generates images. Poems. Movies. Images (AI and not!). Spoken Word. Even a cross-stitch work (can you guess who made it?).

All of which still had a human “in the box.” A story to be told. A meaningful application of whatever medium they were using. Some of them used AI, sure, but they were still involved in the process of making these submissions meaningfully. That, to me, is where the magic is. When we treat humans like they matter, and they have the ability to express themselves and do what they do best: create.

People want to criticize AI more

You might be wondering, given the 22 submissions and the cash prizes involved, who won?

In an interesting upset for a group literally including “AI Interest” in the name, two of the three submissions that won were both a) not created with any AI in the process and b) critical of AI’s implementation one way or another. By a wide margin. To share all three winning contributions:

Winning Submissions

1st Place, Chris Pruneski (Insta: @memento_vivre_arts)
“The Four Horsemen of the AI-Pocalypse”, Acrylic on Canvas.

2nd Place, Fiona Nugent
“AI Poem,” Inspired by the rapid evolution of AI tools.

Third Place, Albert Yu
“HERMES,” A foreign language tool that uses AI to source words from the rarest languages in accordance to one's interests. No pictures or extra links to provide, sadly!

I think the attendees did a kick-ass job with voting based on what actually excited them, made them feel something, instead of just the “best AI powered thing.” The submissions were amazing as I said before, but the three particular winning submissions took the cake (and the bread from our generous sponsors).

I also feel like this spread of winners really says something about what people actually want to hear in this discussion. They saw the strong critique, the unadulterated concern, the meaningful use and discussion of this technology and said THESE are winning submissions. Even with an AI-powered application, there was still a meaningful application and process involving a human. Not replacement, or cutting corners, or saving money on an already insurmountably large profit margin.

It is worth noting that one of our submissions was anonymous as well. I know them, and they’re a prominent voice in the Colorado AI industry, and they felt there was a concern with submitting with their name on it. The submission itself was extremely critical of how AI is being implemented. And they did it anyway! Because they felt so strongly that it needed to be said.

So what does all of this tell you about how people actually feel about AI? It tells me that there’s a gap between the perceived and reality. There’s a number of people with growing concern and frustration, but they don’t feel comfortable bringing things up for fear of retaliation and being seen as backwards. We should be anything but proud of that.

So What’s Next?

Not sure, to be honest. Part of me wants to make these types of events a regular occurrence, but I want to do so while sticking true to my own principles of how I utilize AI. It will take me time to iterate and find a happy medium on that in the context of AI and Art.

To summarize, here’s what I learned from AI and the Creative Arts:

• AI and Art is not Black and White. When you actually involve the creatives affected by this technology, they want to be a part of the conversation. Of course people don’t want to use technology that doesn’t actually help them. And, obviously, the whole ‘us vs. them’ of AI is clearly starting to get old.

• Human-first AI is where the magic happens. All the submissions were made without just pawning off the whole process to AI. Such a variety of beautiful expression of what something means to an individual only happens with a human in the mix. I’m positive we can apply this across disciplines.

• People want to criticize AI more. A great deal of submissions, winning or otherwise, were critical of the current direction of the technology. When people lift up submissions that say AI is a concern, and people don’t feel comfortable attaching their name to critical feedback, and it shows a critical gap between companies and the people they support. If nobody cared, they wouldn’t both with their concerns.

This event proved a vital truth to me, in that there isn't a monolith of singular thought about AI. So, to the prominent individual who dismissed the genuine concerns of our (winning) submissions as 'alarmist' to our creatives’ faces: know that stifling diverse voices is never a good look for any community leader, especially one claiming to build an inclusive one. Step out of your echo chamber and do better, preferably once you’ve cleaned the egg off your face.

Technology, particularly AI, thrives on openness and transparency, not on forced consensus or the suppression of critical perspectives. We have enough oppression to navigate in this world; a truly forward-thinking future for AI must embrace, rather than fear, the full spectrum of human expression, including our concerns and our critiques. People who don’t accept that are truly the ones that will be left behind, not the amazing folks willing to speak up and say something.

On a lighter note, look how long this dog is!