Artificial Intelligence and Contractual Liability Limitations: A Natural Combination?

Maarten Herbosch,Hannes Claes
DOI: https://doi.org/10.54648/erpl2023027
2023-08-01
European Review of Private Law
Abstract:Artificial intelligence (AI) systems are being used increasingly often in everyday contract performance. At the same time, this novel technology and its autonomy present some unique challenges for our traditional legal system. This is related to the fact that while these systems may usually provide excellent results, they are also prone to make manifest mistakes that no human being would make. In this contribution, we examine some of these challenges relating to the contractual liability for the use of AI systems. To this end, we distinguish between two potential ways AI systems may be implemented in a contract. First, the AI system may constitute (part of) the object of the contract itself. Second, an AI system may be used to perform an existing contractual obligation, that does not detail the use of such a system. Both of these types of AI system application lead to unique challenges relating to the contractual liability of the AI system provider or user. In this regard, it is essential to know the extent to which liability may be contractually limited or excluded. The two identified functions AI systems may perform regarding the contract also turn out to be essential in this regard. Put briefly, while liability limitations regarding occasional, abhorrent mistakes appear to be generally permissible when the AI system is the object of the contract itself, they appear to be unacceptable when the AI system is used to perform an existing contractual obligation that is determined without reference to the use of such a system.
What problem does this paper attempt to address?