Overcoming Present Bias in Engineering Design and Construction with Future Thinking and Generative AI

Abstract

Present bias—the tendency to favor immediate gains over long-term benefits—can negatively affect design decisions in construction engineering. Designers often prioritize short-term economic gains that compromises the resilience of the asset, leading to increased cost of remediation in the future. This dissertation explores how mental visualization through future thinking and the use of generative AI tools can help reduce present bias during early-stage design tasks. Three experimental conditions were tested: present thinking (control), future thinking, and AI-assisted future thinking. Civil engineering students (n = 90) participated in constraints identification and concept design tasks for a campus redevelopment project, while their verbal responses and brain activity were recorded. Functional near-infrared spectroscopy (fNIRS) was used to measure cognitive load. To analyze design narrative, qualitative coding and natural language processing (NLP) techniques such as semantic similarity and text network analysis were used. Results show that future thinking and AI assistance improved the quality and future orientation of design outputs. The AI-assisted group identified more climate-related risks, demonstrated higher alignment with futureproofing concepts, and showed more coherent design narratives. These improvements were achieved with reduced cognitive load. Notably, the influence of AI assistance extended beyond the phase in which it was used and enhanced performance in subsequent design stage. The findings support the role of AI as a cognitive support tool that can enhance design thinking, reduce cognitive load, and lead to more resilient and sustainable design outcomes in construction engineering.