Artificial intelligence as a tool for renegotiation

Abstract

This paper investigates the legal and practical feasibility of entrusting an AI system with a role of third- party decision-maker in contract renegotiation by considering the implications under both the Italian Civil Code and the EU AI Act, as well as the broader question of whether AI can truly replicate the role traditionally performed by a human arbitrator. In the context of long-term contracts, unexpected factual or legal circumstances arising after the contract's formation may impair its synallagmatic balance. To reduce such effects, parties may mitigate the imbalanced contract recurring to the renegotiation. The renegotiation of contractual terms can also be entrusted to a third party, based on the principle of contractual autonomy, as provided by Article 1349 of the Italian Civil Code. This provision, according to varying doctrinal interpretations, draws its legitimacy from principles such as contractual good faith (Article 1175) or equity (Article 1374). Thus, Article 1349 not only serves as the legal basis for initially determining the contract's object but also authorizes delegating its redefinition during the contract's term. The text of Article 1349 establishes two approaches: the content can be determined at the third party’s sole discretion or based on their equitable assessment. In the first scenario, if the third party fails to make a decision, the contract is void. Moreover, an erroneous decision by the third party cannot be challenged unless made in bad faith. Conversely, where equitable assessment is involved, the failure of the third party to decide transfers the matter to the courts. An innovative and potentially groundbreaking solution discussed in this abstract involves entrusting the role of third-party arbitrator to an algorithm or artificial intelligence (AI) system. Such a mechanism could redefine contractual terms more effectively by considering the parties expressed interests. This AI-driven system would analyse the interests and actions of the parties involved, determining the appropriate contractual content. This work aims to analyse the various challenges this tool would face. These will be examined in light of the recently introduced EU AI Act (Regulation 2024/1689), which provides further guidance. The law strengthens the possibility of supporting such a mechanism. Nevertheless, entrusting contract renegotiation to AI raises questions concerning its legality and permissible limits. Regarding the methodological approach adopted in this work, the sources used in this paper were identified through a progressive reading process, starting from the works of the most authoritative scholars in the field. The references cited in those foundational texts served as a guide to explore the broader debate and to select further contributions that align with the author’s critical perspective.