Rethinking cars for sustainable mobility. How large language models can leverage change

Abstract

Autonomous driving cars, powered by advancements in artificial intelligence (AI), sensor technology, and enhanced communication capabilities of 5G, are set to revolutionise transportation, promising significant improvements in safety, efficiency, and accessibility. The transition to fully autonomous vehicles should align with the shift to Society 5.0, where vehicles are zero-emission and fully integrated into a circular economy. This shift requires a radical change not only in the automotive industry but also among car users, who are moving from being car owners to car users with shared autonomous electric vehicles becoming a more sustainable public mobility service. This transition is a complex endeavour in a socio-technical system and needs a coordinated effort from many different stakeholders. We conducted qualitative research that included a human survey and prompts for Large Language Models (LLMs) focused on sustainable mobility. This human-focused survey comprised questions about participant expertise, public-private partnerships, policies, stakeholders, consumers, and standards. We carefully crafted the prompts for the LLMs to elicit more accurate, relevant, and contextually appropriate responses. Based on the insights gained from both the final human responses and those generated by the LLMs, we proposed a hybrid methodology that integrates findings from both approaches. This hybrid methodology combines insights, reflecting the current literature and representing an integrated view among all stakeholders to achieve the transition to SAEV and CE implementation. This approach could serve as a reference process for combining LLM-generated responses with real-life human expertise to collaboratively conduct questionnaires, and extend qualitative research methods, especially for complex domains with many different stakeholder interests. Our findings reveal that LLMs offer scalability and speed, complementing human expertise which is often difficult to access and limited in speed. We provide a particular example to illustrate similarities and differences in the process steps and combine the strengths and weaknesses of experts and LLG-generated responses to leverage the insights from different stakeholder perspectives and accelerate the transition toward more sustainable mobility.