Abstract¹

The uptake of Artificial Intelligence (AI) innovations in the construction industry implies new physical and socio-political risks and regulatory challenges. The regulatory analysis is challenging due to the inherent complexity and a lack of applied knowledge specific to AI use for construction activities. This research extends the Normal Accident Theory (NAT) and High-Reliability Organisation (HRO) theory to the AI-fied construction space. I posit that AI-fied construction² can move towards a culture of HRO by regulating precautionarily.

First, to identify the research frontier, I analysed 974 journal articles via a mixed-method approach using keyword co-occurrences followed by activity mapping. I identified the expert technology expectation using a risk workshop with 9 expe...[ Read more ]

Abstract¹

The uptake of Artificial Intelligence (AI) innovations in the construction industry implies new physical and socio-political risks and regulatory challenges. The regulatory analysis is challenging due to the inherent complexity and a lack of applied knowledge specific to AI use for construction activities. This research extends the Normal Accident Theory (NAT) and High-Reliability Organisation (HRO) theory to the AI-fied construction space. I posit that AI-fied construction² can move towards a culture of HRO by regulating precautionarily.

First, to identify the research frontier, I analysed 974 journal articles via a mixed-method approach using keyword co-occurrences followed by activity mapping. I identified the expert technology expectation using a risk workshop with 9 experts from academia, industry, and law. I interviewed 51 construction personnel to obtain the laymen technology expectation. I then identified the potential risks associated with these applications by forming a risk map based on an analysis of 94 articles discussing the ethical aspects of AI. Risks are identified by expert elicitation and interview with 51 construction personnel and 18 personnel from a manufacturing firm. The risks are used to define the regulatory space.

The keyword co-occurrence network and activity mapping indicate the innovation trend towards back-end construction activities. Expert elicitation indicates the development of technologies that bridges the gap between back-end and on-site construction activities. Construction personnel expect a radical change in the construction processes in a short time frame. The risks due to AI systems are beyond physical and includes retrenchment/ layoff, privacy risks and liability risks. The regulatory space is spanned by the exploitation of highly qualified and skilled employees, AI system developers are outside the liability spectrum causing irresponsible innovation, privacy of the client is at risk etc. The country-level laws, ISO standards and contractual guidelines are inadequate for an ‘AI-fied’ construction sector.

¹ A part of this abstract is taken from Pillai, V. S., & Matus, K. J. M. (2020). ‘Towards a responsible integration of artificial intelligence technology in the construction sector’, Science and Public Policy, 47/5: 689–704 with the permission of the authors.

² AI-fied construction meaning, the construction sector that has undergone a transformation due to AI inclusion