Friday 31 May, 12:00 – 13:00 | Online only
Title: ‘Responsibility for Hallucinating: Contractual Devices for Limiting the Scope of Responsibility for Professional Advice and Assistance provided by Generative AI’
Abstract
The emergence of relatively accessible generative AI to create text, images or sounds has sparked a renewed interest in opportunities for using this technology to provide professional information and assistance to prospective clients. Start-ups and established entities are exploring the possibility of refining generative AI to provide personalised information relevant to medical, legal and dispute-resolution services. There are considerable possible benefits for consumers in these developments. However, a defining characteristic of generative AI is that it generates content and produces outputs that are not explicitly programmed, and which occasionally or even commonly is inaccurate or wrong. Clearly the consequences for individual clients may be catastrophic if a generative AI ‘hallucinates’ critical information in providing advice or assistance. While there are discussions globally about the case for specific AI regulation, there is a general assumption that business-to-consumer uses of generative AI will be sufficiently regulated through existing contract, tort and consumer protection laws. This paper explores that assumption, focusing on the efficacy of disclaimers in protecting professional service providers of chatbot advice services from liability in the event of error or inaccuracy.
Bio
Jeannie Marie Paterson is a Professor of Law and the Director of the Centre for AI and Digital Ethics at the University of Melbourne. Jeannie’s research focuses on consumer protection law, especially in the context of emerging digital technologies. She has written extensively on issues of fairness, privacy and existential risk in the emergence of AI and social robots, and consults regularly on these issues to government and civil society organisations. Her current research focuses on chatbots in B2C and in home contexts.