This article examines the validity of contracts concluded through AI-based conversational agents from the perspective of Romanian civil law.
Two main scenarios are considered: the AI agent as a passive communication interface and the AI agent used for recommendations or limited negotiation.
I. Introduction
Artificial intelligence has become an increasingly common interface in commercial relationships. Digital platforms may integrate conversational agents that centralize information, respond to users and facilitate contract formation within predefined limits.
II. Offers communicated by an AI agent
When the agent merely communicates a trader’s predefined offer, the legal analysis remains close to classic e-commerce mechanisms. The key point is that the offer must include the necessary elements and reflect the trader’s intent to be bound upon acceptance.
III. Negotiation, recommendations and limits
The issue becomes more nuanced when the AI agent recommends products or responds to requests to modify the offer. A prudent approach requires clear rules, transparency and predefined contractual instructions.
IV. Conclusion
As long as the AI agent acts within predefined instructions and does not display its own legal autonomy, its use does not change the substance of contract formation; it changes the technological interface through which the trader communicates.