The Adoption of AI in the Construction Industry During the AX Era
▲ Senior Researcher, Won Ji-sun, Department of Future & Smart Construction Research, KICT
Prologue
Following the release of ChatGPT, which surpassed 100 million users within just two months, AI smartphones with on-device AI and generative AI capabilities are also causing artificial intelligence, which once felt like a technology that was far off in the future, to permeate our daily lives. Beyond daily life, applying AI to business is becoming a necessity rather than an option. Whenever new AI technologies emerge, we find ourselves constantly considering how to apply them to our work, and how to formulate strategies for the future. We are living in an era in which we must continually consider the implications of AI advancements for our professions. As AI technology progresses rapidly, the AI Transformation (AX) is becoming a present-day issue rather than a future concern (ETNews, 2024). In the upcoming AX era, what tasks do construction industry professionals wish to apply AI technology to, and what difficulties are they facing in the adoption process? The Korea Institute of Civil Engineering and Building Technology (KICT) conducted a survey in 2022 to gauge the industry's perception and demand for the adoption of construction AI as part of its major project (Research on Smart Construction Technologies to Drive the Future Construction Industry and Create New Markets, 2022-2023). Although the survey results may not accurately reflect the current demand due to changes in the environment, we hope that sharing some of its findings will help in setting future directions and determining appropriate responses to the technology.
Survey Overview and Respondent Characteristics
The overall survey items covered ① the current status and plans for AI adoption at the respondent's affiliated organization, ② perceptions of AI adoption in the construction sector, ③ demand for AI adoption in the construction sector, and ④ barriers to AI adoption in the construction field and measures for creating an ecosystem. In this article, we focused on analyzing item ③, which surveyed the construction tasks that respondents wanted to prioritize for AI technology adoption, and item ④, which examined the long-term measures needed in the construction industry. Items ① and ② were excluded, as we expect the answers to items ① and ② to change significantly depending on the AI market situation. The survey was conducted targeting workers in the construction industry, and was participated in by a total of 107 respondents. Regarding having experience utilizing AI technologies for construction work, 49.5% had such experience while 50.5% did not, an almost equal ratio. In terms of the respondents' affiliated organizations, 29% were from design firms, 22.4% from corporations/public corporations, and 16.8% from academia/research institutes. Regarding job responsibilities, design and construction work accounted for 32.8% and 21.5%, respectively, together representing more than half of the respondents. Approximately 82% of the respondents had more than 10 years of experience in the construction industry, and the facility areas they were in charge of were buildings (42.1%) and roads (34.6%) respectively, together comprising a significant portion.
Current Demand for AI Adoption in the Construction Sector
To assess the demand for AI adoption in the construction sector, we provided a list of tasks in each construction phase where AI could be applied, along with examples of AI applications for those tasks. Respondents were asked to select the tasks they considered most urgent for AI technology adoption, in the order of priority. In this article, statistics on the top priority tasks and the results of a demand analysis according to respondents' characteristics are selectively explained.
Planning and Design Phase
The demand in the planning and design phase was surveyed based on the eight tasks shown in Figure 2. A comparison of demand between all respondents and those in charge of planning/design tasks, as well as the results of a demand analysis according to whether they have experience utilizing AI or not, is as follows:
Both the group of all respondents and the group in charge of planning/design tasks showed high demand for "design analysis and interpretation" to derive optimal design solutions and extract design characteristics, "duration and cost estimation" to predict approximate estimates, and "design planning and plan establishment," like generating various design alternatives. The two groups assigned the same priorities to 8 specific tasks. Comparing the demand based on AI utilization experience, the AI-experienced group showed a noticeably higher demand for "design planning and plan establishment" than the AI-inexperienced group. This is likely due to their practical experience with AI-based design automation solutions, reflecting higher expectations around the benefits of adoption.
Construction Phase
The demand in the construction phase was surveyed based on the five tasks shown in Figure 3. A comparison of demand between all respondents and those in charge of construction tasks, as well as the results of an analysis of demand according to years of service in the construction field, is as follows:
All respondents and the construction task group showed high demand for the adoption of AI in "safety management," such as accident prediction and disaster case classification. and "process management," like process optimization. Notably, the construction task group showed about 15% higher demand for AI adoption in "safety management" than the average of all respondents. This trend is attributed to the increasing importance of construction site safety, highlighted by laws such as the Serious Accident Punishment Act, leading to a heightened perceived need for AI-based safety management technologies in the field. When we compare demand based on years of service in the construction industry, those with less than 5 years of experience showed a relatively higher demand for AI adoption in "quality management," while those with 5 to 10 years of experience showed a higher demand for AI adoption in "progress management" compared to other groups.
Maintenance Phase
The demand at the maintenance phase was surveyed based on four tasks depicted in Figure 4. The comparison of demand for AI adoption in their tasks between all respondents and maintenance task personnel, as well as the analysis of demand for AI adoption in their tasks based on years of service in the construction industry and whether they have experience using AI, are as follows: Both all respondents and the maintenance task group commonly identified "inspection and diagnosis," which deals with damage detection and condition grade assessment prediction, as the most urgent task for AI technology adoption. The maintenance task group showed a higher demand for AI adoption in "repair and reinforcement," which predicts repair methods, costs, and timing, compared to "continuous monitoring,” such as structural condition change monitoring, indicating a difference in perspective between the groups. The demand for "preventive maintenance," such as creating deterioration models and predicting aging, was the lowest. Looking at the demand for AI technology adoption according to years of service in the construction industry, the group with more than 15 years of experience showed the highest demand for "continuous monitoring." Interestingly, the group with 5-10 years of experience showed a higher demand for AI technology adoption in "preventive maintenance" compared to "repair and reinforcement." Examining the technology demand based on AI technology utilization experience, those with AI experience responded that "continuous monitoring" was the most urgent task area for AI adoption, while those without experience said "inspection and diagnosis" was the most urgent area for AI adoption (Won Ji-seon, 2024).
Barriers to AI Adoption in Construction and Measures to Create an Ecosystem
To assess the barriers to introducing AI in the construction sector and prepare measures for facilitating AI adoption in the future, opinions were surveyed by dividing respondents into groups with and without AI technology utilization experience. The difficulties in introducing and utilizing AI were surveyed in the order of "data acquisition and quality issues," "lack of AI-related personnel," and "lack of construction-specific foundational technologies" for both all respondents and the AI utilization experience group. A survey on measures to overcome AI adoption barriers and promote AI adoption in the construction field that was conducted on a group with AI development experience revealed many opinions on the "establishment of AI infrastructure, including data openness.” In addition, tasks such as "nurturing AI personnel," "expanding awareness of AI utilization," "improvement of regulations and establishment of regulatory systems" and "support for AI-related R&D" were identified (Shin Jae-yeong et al., 2023).
Epilogue
This article examined the current demand for construction tasks requiring the introduction of AI technology based on the opinions of 107 construction industry workers. Today, the construction industry is facing new changes with the emergence of generative AI. It is said that the future of technology is determined by how familiar and useful it is to people rather than its innovativeness. To utilize AI technology valuably and usefully in business, it is necessary to first identify the tasks with which help is required, and which problems to solve. We hope this data will help construction industry workers understand their needs and devise their own strategies.
―――――――――――――――――
References
• Let’s Lead the AI Transformation (AX) Era, ETNews (January 1, 2024), https://www.etnews.com/20240101000072.
• Survey Report on Perception, Demand, and Ecosystem Creation Measures for AI Adoption in the Construction Industry, KICT (2022)
• Current Perception and Research Trends of AI in Facility Maintenance, Won Ji-seon (2024), KACEM News, Vol. 242.
• A Study on the Perception of Practitioners for Facilitating AI in the South Korean Construction Industry, Shin Jae-yeong, Won Ji-seon (2023), Journal of the Korea Academia-Industrial Cooperation Society, Vol. 24, No. 6, pp. 386-399.
