Linking generative AI and morphological analysis to conduct foresight evaluation

Purpose This paper aims to demonstrate a new methodology for retrospectively evaluating certain foresight studies. Specifically, this study proposes using Generative AI (Gen AI) to evaluate foresight studies that modeled future outcomes using morphological analysis (MA). Design/methodology/approach This paper demonstrates the new methodology by evaluating the Hudson Institute’s 1976 report, The Next 200 Years: A Scenario for America and the World. This report used MA to systematically pair four different views of the future, ranging from pessimistic to optimistic, against ten different issues. The author used ChatGPT 4o to perform a vertical analysis to determine which of the four views best matched the current world state. The author then conducted a horizontal analysis to validate the first analysis and quantify the fit of each issue. Findings The findings suggest that certain foresight studies using MA can be evaluated quickly using Gen AI. In particular, ChatGPT 4o identified the Hudson Institute’s second-most-pessimistic scenario as the best match to the current world state. A horizontal analysis confirmed the second-most-pessimistic scenario was the best fit at 88%; the most optimistic scenario was the worst fit at 40%. Originality/value The author believes this work has novel implications for foresight evaluation, a relatively undeveloped subfield of future studies. First, evaluating other historical foresight reports using the methodology can potentially help develop a body of knowledge for evaluators. Second, given the ease of evaluation using these methods, the author believes foresight products structured using MA will allow future AI systems to more readily interpret them.