While AI can be harnessed to improve safety and accountability within architecture and construction, we are only just beginning to comprehend the new risks and uncertainties it introduces. Amid all of AI’s potential positive contributions, one question stands out: who is responsible if AI gets it wrong? If an AI tool indicates that an architectural design is compliant with UK regulations, but that later turns out to be incorrect, does liability fall on the software user, the software developer, or the insurer? This case study delves into this very question, shedding light on the emerging challenges.