Creative Algorithmic Intelligence: Capabilities & Complementarity​

A key challenge for the 21st century is enabling constructive synergies between human and artificial intelligence. This challenge arises at the juncture of two major societal trends: the integration of algorithms in most aspects of everyday life, and accelerating technological developments in machine learning (ML).

To facilitate positive synergies between humans and AI, we must appreciate the affordances of each. Notably, creative intelligence is often positioned as the most categorically “human” faculty. Creativity is the ability to generate ideas or artifacts that are both novel and valuable. Machine creative outputs have historically been neither, but machine intelligence is likely to transform creative endeavors. A nascent example is the use of deep learning to style the short film “Come Swim.”

Our project investigates the hypothesis that Creative Artificial Intelligence (CAI) can help unlock creative human potential, enabling forms of creativity unachievable by humans alone. Bringing together a truly cross-disciplinary research team (machine learning, sociology, education) allows us to:

1) Generate rich qualitative evidence identifying the current capabilities of CAI and documenting human experiences of such partnerships, and

2) Provide a robust quantitative framework predicting the future scope and richness of algorithmic creativity.

We will do so through methodological innovation, integrating novel machine learning techniques with elicited qualitative data and analysis in combination with secondary analysis of existing structured datasets. Ultimately, this allows us to address an overarching objective of aiding human flourishing through better understanding human-led creative endeavors assisted by algorithms. This project also delivers scientific progress in ML and aids social science theory building related to AI, cultural production, creativity, and complementarity.