MMOOZZGG.com

AI for lawyers

AI code of ethics in Russia: a guide for lawyers and companies

5 min read6 sources

The AI code of ethics was adopted in Russia on 26 October 2021 and is still a recommended document: joining it is voluntary. Below is what its text says, which clauses are worth moving into company policy and contracts, and how the code relates to 243-FZ, Russia’s AI law.

On this page · 4 sections

What the code is

The code sets general ethical principles and standards of behavior for participants in AI relations, whom the document calls AI actors. It covers the ethical side of creating, introducing and using AI at every stage of the life cycle, where these questions are not settled by law or by technical regulation.

The document has two sections: “Principles of ethics and rules of conduct” and “Application of the code”. The recommendations apply only to civilian AI systems, and industry and internal documents can extend and specify its provisions.

The circle of actors in item 1.3 of section 2 is wide: developers, customers, data suppliers, experts, manufacturers, operators of systems and those who take part in regulation, including foreign ones. A law firm that uses AI falls into it as a customer or an operator.

Joining is voluntary (item 2.1 of section 2). An organization that joins appoints an AI ethics officer, if possible at once or within two months (item 2.2), and the register of members is public (item 2.4).

Main principles

Section 1 is built around six ideas: protection of human rights, responsibility in creating and using AI, human responsibility for the consequences, use of AI for its intended purpose, the interests of technology development above the interests of competition, and transparency in informing about the capabilities and risks of AI.

The clauses a lawyer needs most often:

1. A human-centered approach (item 1.1): human rights and freedoms are named the highest value; this is the code’s main ethical principle.

2. Compliance with the law (item 1.3): actors know and follow the law of the Russian Federation, including on legal liability.

3. AI identification (item 2.5): the user is honestly told that they are talking to an AI system when this affects human rights and critical areas of life, and they get a way to stop such a conversation.

4. Safe work with data (item 2.6): follow the law on personal data and on secrets protected by law when you use AI systems.

5. Oversight (item 3.1): human supervision, logs of the system’s work and the ability of a person to cancel legally significant decisions.

6. Responsibility (item 3.2): a person, an individual or a legal entity, always answers for all the consequences of the system’s work.

Item 2.6 matters most for legal practice: secrets protected by law, including attorney-client privilege, are directly part of the ethical duties when working with AI. How this relates to Article 8 of the Russian law on the legal profession is covered on the page “Attorney-client privilege and AI”.

How to refer to it in contracts

A sample; adapt it with your firm’s lawyer.

The code is a recommendation, so the general phrase “the parties follow the code of ethics” gives little in a dispute. Something else works: move specific clauses of the code into the contract as the parties’ duties with a clear subject.

1. The contractor tells system users that they are talking to AI in cases that affect their rights (under item 2.5 of the code).

2. The contractor keeps logs of the system’s work and hands them to the customer on request within a term agreed by the parties (under item 3.1).

3. An employee of the customer makes and signs legally significant decisions; the system prepares the draft and the materials (under items 3.1 and 3.2).

4. The contractor processes personal data and information that is a secret protected by law under the law of the Russian Federation (under item 2.6).

5. The contractor uses the system for its stated purpose and only for the tasks listed in the annex to the contract (under item 4.1).

The same clauses fit internal documents. The code itself allows internal documents that extend and specify its provisions, so a company AI usage policy can rely on it directly.

If the counterparty has joined the code, write its AI ethics officer into the contract as the contact for disputed questions about using the system (item 2.2 of section 2). You can check membership in the public register.

Link to 243-FZ

The code is five years older than the law. Federal Law No. 243-FZ of 26 July 2026 took effect on 1 September 2026 and regulates large foundation models, while the code limits itself to questions the law does not settle.

The principles largely match. Article 4 of the law lists among the principles of regulation the protection of human rights and lawful interests, respect for human autonomy and free will, a risk-based approach and safety; in the code, the same ideas sit in items 1.1, 1.2 and 2.1.

The difference is in whether it binds. Participants answer for breaking 243-FZ under the law of the Russian Federation (Article 11), while the code has no sanctions. The duties of developers of sovereign and national models and labeling under Article 13 of the law take effect on 1 March 2027.

The practical conclusion: until 2027, the code gives ready language for policies and contracts, and 243-FZ sets the frame these documents will have to be tied to. A general map of the rules is on the page “Legal regulation of AI”.

Sources and fact-checking

Checked by the editors on 16 October 2026.

  1. AdIndex, 26.10.2021 (adindex.ru)
  2. AI code of ethics, official text, 26 October 2021 (ethics.a-ai.ru)
  3. ppt.ru, 26.07.2026 (ppt.ru)
  4. ConsultantPlus, 243-FZ, Article 4, 26 July 2026 (consultant.ru)
  5. ConsultantPlus, 243-FZ, Article 11, 26 July 2026 (consultant.ru)
  6. ConsultantPlus, 243-FZ, Article 13, 26 July 2026 (consultant.ru)

Found a mistake? Tell us

Publications

Contact us