Computer-aided design tools for superminds: Understanding user needs and evaluating design options

Abstract

The rise of machine learning and progress on the path towards advanced artificial intelligence requires us to think about the future of work, design and how both will be conducted when humans and machines work together. The supermind design methodology was developed to aid the formation of groups of humans and machines to solve problems together.

This thesis addresses the future of collective human-computer problem solving at two levels. The first is to identify the best initial user segment and user needs for “computer-aided design tools” to help users apply the supermind design methodology. The second is to design the feature set, user interface and ongoing user analytics for a software tool powered by the GPT-3 deep learning model that can support practitioners of the supermind design methodology.