Assisted Volition: Mistake, Fraud, Influence, and the Allocation of Remedies in AI-Assisted Decisions under Argentine Private Law

Artificial intelligence can supply the information, comparisons, and formulations through which a person decides to enter a legal transaction. A subsequent signature establishes neither that the person independently assessed those materials nor that their influence invalidates the transaction. This article develops a doctrinal account of assisted volition under Argentine private law. It distinguishes informational assistance, deliberative recommendation, expression or transmission, and authorized execution. Its central proposal is a double reconstruction: first, a constitutive reconstruction of the act, declaration, parties, and authorization; second, an epistemic reconstruction of the alleged defective premise, its provenance, and its contribution to the decision. That sequence permits a differentiated application of factual mistake, fraud, threats, lesion, representation, and the rules governing invalidity and damages. Particular attention is given to counterparty recognizability, human and organizational knowledge, nondeceptive preference formation, subsequent approval, evidentiary asymmetry, and contractual allocation of automation risks. A bounded comparison with automated-contracting instruments and Singaporean case law identifies attribution problems without importing foreign validity rules. The article argues that the validity of the transaction, the performance of the assistance service, and the distribution of resulting losses must remain separate inquiries. It offers a structured method for investigating their interaction, rather than a general presumption of defective consent or a new category of machine intention.