ABSTRACT
As AI programs are becoming sophisticated in contract negotiation, this paper primarily focuses on AI-negotiated contracts and the challenges they bring to the traditional contract law doctrine. It further explores how concepts around intent, offer, and acceptance handle the novel challenges AI programs come with. Building on this, the paper raises another important question around mistake doctrine and whether it can handle situations when AI systems, when negotiating on behalf of their deployers, produce undesired outcomes, and who should bear responsibility for this eventually.
Although extensive scholarship exists around agency law, the objective theory of contracts, and the mistake doctrine as potential frameworks for addressing these challenges AI programs bring, it has not yet fully accounted for the evidentiary opacity of modern machine learning systems or the cross-jurisdictional complexity of AI-negotiated agreements. Thus, this paper also brings some analysis from a comparative approach between civil law jurisdictions and common law jurisdictions, drawing parallels on where they have shared features and where they show divergences. In light of these nuances, the paper argues that, for the most part, the challenges arise around two main aspects, which are allocation of risk in AI systems and how evidentiary issues will be handled across jurisdictions.
Omonov, Sherzodbek, Who Clicked Accept? Algorithmic Error, Evidentiary Opacity, and the Allocation of Contractual Risk Across Legal Systems (May 8, 2026).
Leave a Reply