The sessions at the fifth annual Creativity Conference at Southern Oregon University triggered the following perceptions and ideas. I will strive to cite specific presenters, when possible, given my, at times, nearly illegible notes.
In general:
The sudden ascension of AI creation bots challenges current definitions of creativity and provides new tools for investigating creativity in addition to providing new creative artifacts. People are devising clever ways to measure creativity in action. Working on creative endeavors is empowering. It was a joy to spend time amidst all that focused, intent, intelligent activity.
Generative Artificial Intelligence (GAI) creates new things:
Creativity entails newness, surprise and utility. New to whom? It must be new to humans. Can something be new to a bot which can generate newness at will? Using the term “at will,” in relation to a bot is just weird. Only humans can be surprised, though bots may be able to act surprised. Yet, if a bot routinely creates “surprising” artifacts are they any longer surprising? Will we be consigned to remarking, “Oh, another bot generated great idea. Ho-hum.” ? At this point, AI bots are constructing objects useful to humans, essentially because the humans asked for them. But it is possible to think of asking a GAI bot to devise a tool that would allow it to do its work better. Or a meta thought: asking the bot to ask itself routinely to improve its processes.
Humans will need to respond carefully to all these new creative objects:
If humans are going to have any control over these AI bots, they will have to become better reviewers, judges and raters of creative artifacts. In music and art, we speak of a work evoking emotions and even touching the soul. Humans will need to learn to specify exactly what characteristics produce those outcomes, as reductive as that seems. This may be a way of separating human-created objects from AI- generated objects. Unfortunately, GAI can spam us with thousands of new creations in response to a single prompt. Human raters will need to have firm background in the space of a particular creation, though it might be possible to have an AI bot do the background search. In any case, humans will need to judge the ethical content of AI artifacts. This step will be necessary and people will need to be trained in this type of review. Thus, an imperative to demand careful critical thinking and precision from these GAI’s, which means that humans will need to be better and more precise thinkers – like parsing a mathematical proof. Knowledge of how to ask questions of an AI bot will be an essential skill.
The role of humans amidst all these new creative objects:
In any group endeavor, there is a need to communicate one’s thinking and to recognize that humans think in different ways. If an AI bot is part of a group, it will need to communicate its thinking in different ways for different people. The group facilitator may need to prompt the bot to explain itself in different ways. Often human beings are persuaded by a story, will GAI have this mode of argumentation? Should we treat AI bots as a new intelligent species?
What happens when the AI goes out into the real world:
AI can train you. Ai can give you more ideas. If a teacher is a creative actor, then AI can help. Still, we will need highly developed learning skills. So far, the bots are creative in two dimensions but not generally in three. But we live in three (four) dimensions. Right now, there needs to be a translation step when interacting with AI about life in 3D.
Errors in some statistical studies have been found when their results are compared to repeated human experiences. AI results will also butt up against reality eventually. Unfortunately, the feedback loop will take too long. We will have to demand that GAI’s explain their art.
What happens when you use AI as a tool:
AI has access to all backgrounds so it will “know” more than any one of us. Reliance on AI answers may result in a fixation on only those ideas that the AI has. Writing is an essential tool in science. AI bots can ease this burden, so its use needs to be taught and practiced. One should be aware that complex processes often need delicate control often learned through trial and error. AI may not be able to replicate this particularly if the processes are physical. Asking a GAI enough times saturates the layers. The AI may have blind spots like the latest discovery on how to beat a computer Go playing bot. AI will be much more useful when it can work in 3D space.
Math Thoughts 