My Creative AI Co-Worker

Creativity has been cited as the domain of humanity (Pistrui, 2018). It includes problem finding, ideation, and problem solving, higher-level processes that are often valued as inherently human and beyond the repetitive, easily automated tasks typically associated with artificial intelligence (AI). This study, however, works from the somewhat controversial assumption that AI has the ability to complete creative tasks at the human level in some fields. I tested this assumption in my integrative paper, my program’s requirement to advance to candidacy. I used content analysis of O*Net occupations and competencies and publicly available documentation of claims of AI capabilities, and found that capabilities of existing AI tools closely aligned with existing creative skills required for creative occupations. If we do accept this assumption, how can we approach AI creativity from the perspective of augmentation of humans, rather than replacement? Why shouldn’t we increase our creative abilities through the use of AI as a tool? These are foundational questions to this study. In an attempt to begin to answer these questions, I am proposing a study that examines the possibility of using AI as a creative thought partner. This study will use a modified large language model (LLM) such as ChatGPT or Bard that “knows” the state of the human user to introduce subtle nudges in the form of divergent creativity. I will measure if the introduction of the LLM has any bearing on the levels of flexibility or originality of the human user’s defined problems vs. that same user bouncing ideas off a human creative thinking partner.

July, 2022 - Present

Research Areas: Center and Lab Affiliation: