How people think, judge, and create with AI
My research examines how creative agency, judgment, and responsibility operate when generative AI systems participate in the production of ideas. I study how people generate, guide, and evaluate AI-supported creative work across the full Creative Problem Solving process, including framing, generation, refinement, and convergence.
As generative systems make idea production abundant and inexpensive, human expertise increasingly expands to include creative direction, criteria construction, disciplined judgment, and sustained involvement in generating and shaping ideas. My work investigates how these capacities develop and how they can be taught in educational and professional contexts.
Research Questions
I organize my work around four connected questions:
- Creative skill: How do individuals develop the capacity to guide, constrain, and refine AI-generated work as a distinct creative skill?
- Judgment: How does human judgment operate when generative systems produce options at scale?
- Agency: What happens to creative agency when generative systems contribute substantially to production, and how is agency maintained through continued human involvement across the process?
- Ownership: How is creative ownership claimed and sustained when generative work is performed by a system?
Core themes
Across projects, I return to:
- Human–AI co-creativity as a process of co-generation, direction, and judgment, rather than simple tool use
- The role of evaluation, framing, and criteria construction in AI-supported creative work
- Prompting as a form of creative literacy, and where it breaks down
- How assumptions within Creative Problem Solving shift under AI-mediated creative processes
- Motivation, autonomy, and agency-preserving design in AI-supported learning environments
- Creativity education as a site where creative identity and agency are developed
Conceptual Framework
Creative Problem Solving describes creativity as a structured yet flexible process involving clarification, ideation, development, and implementation. Movement among these components is guided by the deliberate alternation between divergence and convergence. Traditionally, the creative process assumes that generating ideas requires substantial cognitive effort and that insight emerges through sustained human engagement with problems, constraints, and domain knowledge.
Generative AI alters the operating conditions of this process by dramatically expanding the number of possible ideas that can be produced at very low cost. Systems can rapidly generate alternatives, elaborate concepts, and propose variations across multiple stages of the creative process. This expansion changes the dynamics of divergence and convergence by making idea generation abundant while increasing the importance of how ideas are generated, interpreted, refined, and directed through ongoing human involvement.
Creativity therefore becomes an iterative collaboration in which human and machine co-generate and shape ideas, rather than a purely human production activity. Within this environment, the human role does not disappear from generation but extends across all phases of the process. Individuals frame problems, generate and transform ideas in collaboration with AI, establish constraints, guide exploration, evaluate outputs, and determine which directions merit further development.
Despite these changes, creativity still depends on the interaction of knowledge, imagination, and evaluation. This relationship was formalized by creativity researcher Ruth B. Noller in the formula C = fₐ(K, I, E), which proposes that creativity emerges from the interaction of knowledge, imagination, and evaluation activated by creative attitude. Generative systems can expand imaginative exploration by producing large numbers of possible ideas, but effective creative work still depends on human knowledge to interpret the problem space, human imagination to participate in generation, and human evaluation to determine which ideas have value. Abundance alone does not produce meaningful outcomes without these human capacities.
As a result, expertise increasingly involves the ability to manage abundance while remaining actively engaged in generating, shaping, and evaluating ideas. Individuals must resist premature convergence, recognize weak or misleading outputs, maintain ownership of decisions, and navigate large volumes of generated material without cognitive overload. These capacities involve metacognitive monitoring, criteria construction, and disciplined judgment throughout the creative process.
In educational contexts, this shift places greater emphasis on developing students’ ability to participate actively across the full creative process in AI-supported environments. Learners must cultivate domain knowledge, engage in idea generation alongside AI systems, develop criteria for quality, and maintain responsibility for the work they produce. Creativity education therefore becomes an important context for developing discernment, creative direction, and agency while preserving human creative capacity within AI-mediated environments.
Featured research (2023)
Integrating Artificial Intelligence into Creativity Education: Developing a Creative Problem-Solving Course for Higher Education (2023)
M.S. Creative Studies, SUNY Buffalo State
Links:
Working notes
I maintain an ongoing set of research notes where I test ideas, define terms, and develop arguments in public.
Collaboration and speaking
I occasionally collaborate with researchers, educators, and institutions working on creativity education, CPS, and human–AI co-creative learning. If you are interested in discussing a research question, a course design problem, or a scholarly exchange, email is the best point of contact.
selma [at] selmadawani [dot] com
This page is a living research overview and will be updated as projects and working papers develop.