Towards human-centered artificial intelligence (AI) in architecture, engineering, and construction (AEC) industry

Abstract

Over the years, AI has been utilized as a powerful tool to address complex challenges in the AEC industry. In the current AI practices, machines predominantly plan, manage, control, and optimize work without appropriately considering human-related input and preferences. However, architects, engineers, managers, clients, and other decision makers should consider their input into their work to better generate their desired ideas, prototypes, and solutions. In addition, significant decisions in the AEC industry are mainly reliant on the heuristic processes where assumptions are developed from past experience. However, the current level of AI is not able to properly deal with such human information and experience. This fact especially in large projects can result in a failure to properly utilize the full benefits of AI. Thus, human-centered AI is an essential need to help the machines understand and utilize human input for amplifying human abilities and reflecting realistic conceptions in the AEC industry. This paper presents the major aspects and applications of human-centered AI in the AEC industry and discusses the anticipated benefits and challenges of this technology. Human-centered AI, mainly via natural language processing and machine reading comprehension, can understand and learn from human interests, preferences, languages, and behaviors for providing human-centered environments, systems, and approaches that satisfy human interests and preferences. As the major benefits, human-centered AI is expected to result in architectural processing optimization, design and engineering capability enhancement, data driven project management, collaboration improvement, and safety enhancement. Personalization of human-centered AI and training its systems are considered as the major challenges in developing this technology in the AEC industry. In addition, AEC-specific guidelines and statements should be regulated in developing human-centered AI utilized in hazardous areas. Human-centered AI is anticipated to provide the highest level of human control in the current fast-growing automation of the AEC industry.