AI Litigation

AI-Related Disputes

How are artificial-intelligence disputes handled in court?

Georgia has no separate procedural statute for this field — the dispute is examined under ordinary civil-procedure rules, and the technical questions are resolved through expertise and written, including electronic, evidence.

Does an electronic document have evidentiary force?

Yes. An electronic document as defined by legislation, as well as a document confirmed by electronic signature or a certified document, has evidentiary force. System logs, correspondence and digitally recorded outputs are therefore material that can be produced in court.

What if the evidence is held by the opponent?

Where a party was unable to obtain written evidence from the person holding it, the party may apply to the court for its production. The applicant must substantiate which circumstance relevant to the case can be established by the evidence and what grounds the assumption that it is in the hands of the person indicated.

Is the expert's conclusion binding on the court?

No. The expert's conclusion is not binding on the court — it is assessed under the general rules for evaluating evidence. A court refusing to follow it must reason its refusal in the decision or a ruling, which gives the party room to defend its position.

Can several experts be appointed?

Yes. Where several experts are appointed they may confer with one another; if they reach a common opinion, all sign a single conclusion, while a dissenting expert draws up a separate conclusion. In technical cases this is often necessary, since assessing one system requires knowledge of different profiles.

5 min·8 Feb 2026

About this service

A dispute connected with artificial intelligence — damage caused by a system, disagreement over a model's outputs, improper use of data or breach of contractual obligations — is resolved in Georgia under the ordinary rules of civil procedure: there is no separate procedural statute for this field. The substance of such a case, however, is technical, and success depends largely on how the parties handle the two decisive instruments: electronic evidence and court expertise. This service is built precisely on those instruments. We assist parties to disputes in the field of artificial intelligence — from framing the claim and systematising the evidence, through formulating the questions for the expert, to assessing the expert's conclusions.

Electronic evidence — where every such case begins

Under Article 134 of the Civil Procedure Code, written evidence comprises acts, documents, and business or personal letters containing information about circumstances relevant to the case. In artificial-intelligence disputes this category is broad: system logs, records of requests and responses, archives of model outputs, correspondence conducted by e-mail — all of it may contain information about circumstances relevant to the case. The law adds an important aid: an electronic document as defined by legislation, as well as a document confirmed by electronic signature or a certified document, has evidentiary force. This means that material recorded digitally does not lose its right to be produced in court merely by being intangible. Written evidence is submitted to the court by the parties; and where a party was unable to obtain the evidence from the person who holds it, that party may apply to the court for its production. The party seeking production must substantiate which circumstance relevant to the case can be established by that evidence, and what grounds its assumption that the evidence is in the hands of the person indicated. A refusal to order production must be reasoned by the court in a ruling.

The appointment of expertise — when it is ordered and on whose initiative

Article 162 of the Civil Procedure Code sets the rule: where a judge lacks special knowledge on a question connected with the examination of a case, the court may, on its own initiative, appoint expertise at any stage of the proceedings — but only where the interpretation of that question is essential for the decision of the case and a ruling is impossible without it. Assessing the behaviour of an artificial-intelligence system is exactly such a question: a judge as a rule has no technical knowledge of a model's architecture, training data or output defects, and the fate of the case often hinges on the expertise. In such a case the court issues a reasoned ruling. The parties also have a route of their own: they may independently arrange for expertise to be conducted, in which case the expert's conclusion must be submitted to the court at the stage of filing or preparing the case. When filing the claim, a party is entitled to request a time limit for producing the conclusion. Producing a conclusion at the main hearing is possible only exceptionally — where the need for it could not have been known at the preparation stage for objective reasons, or where the party for an excusable reason failed to secure it. One further rule matters: failure to produce a conclusion cannot serve as a ground for adjourning the case — the court may only set the party a deadline.

Read more on this topic