Private Law Theory

ABSTRACT
This article completes a trilogy, with two companion studies, examining how private international law should respond to cross-border contracts negotiated and concluded entirely by autonomous AI agents. Having established that such contracts admit no choice of law and, for their contractspecific jurisdictional basis, no choice of forum because both doctrines require a bilateral manifestation that unilateral deployment-stage configuration cannot supply this article asks a third and equally pressing question: once the applicable law and forum are settled, what standard governs the deploying party’s performance, and who bears responsibility when an autonomous agent’s conduct falls short of it? The article argues that the same absence of consent identified in the companion studies is precisely what makes an implied duty the correct doctrinal response: where the UNIDROIT Principles’ duty of best efforts and general good-faith obligation already supply content-neutral standards that do not depend on what the parties actually manifested to one another, autonomous-agent contracts are simply the case those standards were built to reach. It further argues that responsibility for a breach of that implied duty should presumptively fall on the deploying party, not the underlying model’s developer, because the deployer controls release, supervision, and the decision to profit from the agent’s operation. The argument is situated against the European Union’s 2025 withdrawal of its proposed AI Liability Directive, which leaves a fault-based liability gap this article’s contract-based route can partially fill without requiring new legislation, and draws comparative support from the Civil Code of the Republic of Uzbekistan’s own, currently underdeveloped, good-faith and reasonableness principle.

Abdugaffarov, Davronbek, Implied Duties of Care and Deployer Liability for Breach in Contracts Concluded by Autonomous AI Agents (August 9, 2026).