University of the West of England

and

University of Salamanca

The opacity and automation of current artificial intelligence (AI) operations in construction law present a huge challenge for existing legal frameworks. The existing situation is reviewed in this article alongside regulatory and statutory moves in the European Union and the UK. A liability continuum is devised to capture current options for a more purposeful intervention into how multi-party liability can be regulated in construction law scenarios. The abstract model presented categorises the parties considered into three broad groups – Creators, Developers, and Users. The variable approach to liability for these parties reflects increasing levels of opacity and automation. In doing so, the model offers a foundation for future legal approaches to accountability in increasingly autonomous AI construction industry environments.

Artificial intelligence (AI) is having a transformative impact on the contractual and tortious liability landscape in construction projects. Traditional reliance on causation and identifiable human actions and actors is disrupted by AI. AI introduces, amongst other things, opaque decision-making processes, predictive algorithms and automated interventions. Unlike humans, we expect machines to be perfect and not fail. This is a fallacy given that machines are human artefacts and are inherently imperfect.1