Legal awareness: why Russia generalizes judicial practice in cases with AI

The need to summarize judicial practice in cases involving the use of artificial intelligence may be caused by an active discussion of the government's draft law on AI on the eve of its submission to the State Duma, lawyers believe. A large-scale study of this practice was announced on May 27 by the press service of the Armed Forces. Copyright disputes over AI-generated content are increasingly becoming the subject of court review, neural networks are used by parties in courts, and the courts themselves do not have a unified approach to who to hold accountable for mistakes made when using AI. The generalization of practice should help improve the legal regulation of the use of artificial intelligence in Russia, experts believe.

The Supreme Court will summarize the practice in cases with AI

For the first time, the Supreme Court will conduct a large-scale generalization of judicial practice in cases involving the use of artificial intelligence, the press service of the Supreme Court reported on May 27. Its goal will be to create unified approaches to the consideration of offenses and disputes arising in the context of the rapid development of digital technologies.

First of all, the courts will pay attention to issues related to compensation for damage caused by AI decisions and the choice of a defendant in such claims. A separate section of the SUN will be devoted to the protection of copyrights, as well as the protection of honor, dignity and business reputation in case of dissemination of defamatory information, in particular, using deepfakes.

The Supreme Court will analyze cases related to the recording of offenses by video analytics and facial recognition systems; qualification of crimes using deepfakes and fraudulent schemes with chatbots. "The courts will have to evaluate the practice of recognizing documents and conclusions generated by AI as evidence, as well as cases of judicial appeals based on the grounds that the judge used AI tools to make them," the press service of the Supreme Court noted.

Having studied judicial practice, the Supreme Court of the Russian Federation plans to provide the necessary clarifications, and the results of a large-scale analysis can form the basis for future changes in legislation.

The need for generalization is caused by the active discussion of the government's draft law on AI among lawyers and IT specialists, believes Mikhail Bozhor, head of the digital law practice at Afonin, Bozhor and Partners law firm.

— This bill has been criticized by the professional community and the Presidential Codification Council. As a result, some of the provisions have already been excluded from the draft law. Most likely, there will be other changes before the bill is submitted to the State Duma. It was against the background of the controversy surrounding this bill that the Supreme Court was instructed to summarize the practice in cases related to artificial intelligence," the lawyer noted.

Given the fragility of the legal regulation of artificial intelligence in Russia, the conclusions of the Supreme Court will be more likely to be recommendatory, and strict rules will be fixed in the new AI law after its consideration and adoption, Bozhor believes.

Voice communications: the number of AI cases in the courts has increased by 40%

Fraudsters resort to deepfakes, and motorists dispute incorrect fines.

The need to summarize the judicial practice of AI-related cases arose due to legal uncertainty, said Veronika Polyakova, managing partner of the Business Legal Group Bar Association.

"The current legislation does not keep up with what is happening in reality, it does not contain guidelines for judges to resolve cases with AI," she stressed. — There are more and more court cases related to the use of AI every day. And legal uncertainty prevents uniform practice.

Each judge resolves the case as he understands it, within the framework of judicial discretion. As a result, someone's violated right in a similar case is protected, while someone's remains unprotected, the lawyer noted.

Problems with AI

Veronika Polyakova believes that the list of problems related to the use of AI technologies is wide. Among the most pressing issues is who will be responsible for the negative consequences of AI decisions: the user himself, the data provider in a particular field, or the developer.

— In addition, the problem of challenging decisions made on the basis of AI proposals is relevant, — the lawyer added. — The issue of recognizing procedural documents and court decisions written with the help of AI as evidence is also quite acute. It is not clear how to appeal them, and what responsibility is provided for users — interrogators, investigators, prosecutors, judges, and their assistants.

Oleg Bezik, founder and CEO of the Digital Research Laboratory, identified two problems. First, the algorithms are wrong — for example, social networks do not always filter content correctly. Secondly, lawyers thoughtlessly use neural networks without checking the facts.

A recent case confirms the lawyer's words. In May 2026, the arbitration court of the West Siberian District considered it contempt of court for lawyers to file a cassation appeal, which, as it turned out, was completely generated by a neural network. AI made references to non-existent cases in the document and made statements that none of the judges had uttered. The company, on whose behalf unscrupulous lawyers acted, had to answer for this lie.: The court fined her 50,000 rubles.

But there are also reverse examples where the use of a neural network does not affect the legality of a court document.

Prompta and contra: how Microsoft influences the outcome of the trial on the ethics of using AI

Leading neural networks analyzed the chances of concluding a deal between the Pentagon and the company Anthropic

So, the lawyers filed a complaint with the Krasnodar Regional Court — they challenged the verdict of the Yeisk City Court against former policemen Sergei Frolov and Alexei Falun, who were sentenced to seven years in a penal colony for bribery. The defenders were alarmed by the strange wording in the document: one of the defendants was called a "determined official," and they themselves were called "a whole galaxy of defenders." A linguistic examination conducted at the initiative of the lawyers confirmed: The text of the verdict was indeed written using AI. Nevertheless, the judicial board decided that the style did not affect the objectivity of the judge's decision, and left the verdict unchanged.

It is important not to ban AI, but to understand how to handle it in a legal sense, says Vyacheslav Bogatkin, a forensic expert and IT specialist.

— From the point of view of forensic examination, the issue of the digital footprint is becoming key. To evaluate materials created or processed by AI, it is important to record not only the final file, but also the chain of its appearance: initial data, prompta, and so on. Otherwise, a party may bring to court a convincing-looking document, image, audio, or conclusion, but it will be extremely difficult to verify its origin.

AI can help lawyers, experts, and courts, but it should not replace professional assessment, human responsibility, and the verifiability of evidence, say IT experts.