How Russia will regulate AI

The European Union in March adopted the world’s first law regulating the development and use of artificial intelligence (AI). A number of similar initiatives are being discussed in other countries, and they are working on introducing legislation regarding AI in Russia. Lawyers assess the European law as very detailed, but ambiguous, at the same time believing that in Russia the adoption of an AI law is a matter of time. According to market participants, many problems related to artificial intelligence relate to ethics and can be resolved by companies themselves, and regulation should be based on real risks, not phobias.

First Law: Relevance and Criticism

In March, the European Parliament supported by a majority vote the law on artificial intelligence, agreed upon by representatives of 27 EU countries in December 2023. 523 parliamentarians were in favor of adoption, 46 were against, and 49 abstained. Thus, the world’s first law regulating AI was officially adopted. According to EU procedures, it may not come into force until next year.

The law divides all artificial intelligence-based systems according to the level of risk – from low to unacceptable. The document prohibits the use of facial recognition systems and other solutions in the field of remote biometric identification in public places in real time, as well as emotion recognition systems. The police are prohibited from using predictive analytics systems to prevent violations.

Among other things, the law introduces restrictions on the use of generative AI (for example, the ChatGPT neural network) and “high-risk systems based on artificial intelligence,” which include self-driving cars and medical equipment.

“Most likely, after the regulation comes into force, it will become a good example for many other jurisdictions, which can take it as a basis for creating their own rules,” notes Spartak Khulkhachiev, leading legal adviser in the intellectual property practice of the EDB law firm.

The law on regulation of AI in the EU was adopted in a timely manner, when without such an act further technological development could either slow down or create a threat of human rights violations, says Vladislav Eltovsky, head of digital law and intellectual property practice at Seamless Legal. “This law contains perhaps the most detailed regulation of AI based on a risk-based approach and regardless of the method of application of AI,” he said.

The EU law is a breakthrough in the field of AI regulation, says Digital & Analogue Partners lawyer Anna Botvinkina, but a number of experts believe that it does not fully meet the requirements for the protection of human rights. “For example, the law did not completely prohibit the use of emotion recognition technologies and remote biometric identification in real time in public places,” the lawyer clarifies.

Managing Director of the Center for Human-Centered AI of Sberbank Andrey Neznamov also points out the ambiguity of the law. “The ideas of a single horizontal law on AI are present only in a few countries, they are sharply criticized by the technology community. In particular, EU companies spoke out against the AI ​​law last summer. A similar situation occurred in Canada during discussions of a similar act,” said a Sberbank representative.

Boris Yedidin, deputy chairman of the commission on legal support for the digital economy of the Moscow branch of the Russian Bar Association, believes that the legislative practice of the European Union, as a rule, sharply limits the development of new technologies. “As a result, Europe has not yet created its own competitive search engines, audiovisual services, or messengers. The implemented bill of the European Parliament is certainly interesting; it is important to regulate the development of AI, create rules for its use taking into account the degree of potential danger. Some technologies should be developed only under the control of the public or the state – in this part, certain elements of the initiative can be accepted by Russian legislators,” the expert believes.

Practice in the USA and China

US President Joe Biden signed a decree in October last year, which is similar to the European law in terms of requirements for AI system developers regarding transparency of processes. American companies will now be required to provide the government with data on the results of security testing and other critical information before the development goes to market. The National Institute of Standards and Technology has been tasked with developing criteria and the US Department of Commerce with guidelines for labeling AI-generated content. Particular attention is paid to increasing the security of personal data.

In China, last summer, temporary measures were taken to make AI developers responsible for all generated content.

The rules introduced a ban on content that undermines socialist values ​​or incites the overthrow of the government. The developer is responsible for the personal data of users and compliance with intellectual property rights. Also, according to the rules, the need is established to prevent children and adolescents from becoming dependent on the use of generative AI.

Russia is looking for its way

In Russia there is no systemic regulatory framework for regulating AI yet; it is only being prepared. Since last year, United Russia has been working on a bill that is designed to determine the responsibility of developers and eliminate cases of artificial intelligence being used for fraudulent purposes. “Together with ministries, we plan to assess changes in the labor market related to the spread of AI solutions. Now technologies that were supposed to save people from unnecessary physical labor are instead trying to take on the solution of creative problems,” commented Anton Gorelkin, deputy chairman of the State Duma Committee on Information Policy, on the United Russia Telegram channel. The message said that there was no talk of banning neural networks.

Last December, the government introduced a bill to the State Duma obliging AI developers operating in an experimental legal regime to insure liability for possible harm to life, health or property when using the technology. A significant innovation of this initiative is the proposed reduction in the processing time for applications to participate in the EAF. It is not yet clear which regulatory path Russian legislation will take.

Preparation of a bill and its coordination with interested players – government agencies, development companies – can take quite a lot of time, notes Boris Yedidin. “So, it took more than two years to adopt an initiative to regulate related technology – recommendation services. The debate on regulation in the field of big data has been going on for more than five years. This is largely due to the search for a balance, how to protect the rights and interests of the state and users and not harm technological development or limit the emergence of innovation,” the expert comments.

Developers are waiting for a solution

HRlink partner and development director Dmitry Makhlin is confident that the law will appear in Russia within the next one and a half to two years. “The main thing is that it meets the requirements and level of market development,” he emphasizes, adding that, most likely, legislators will focus on the experience of the European Union and China. “The initiative is absolutely timely,” says Dmitry Makhlin. “Artificial intelligence is beginning to move from the stage of mature innovation to the stage of sustainable technology that is used by the mass consumer. If you start regulating the market now, you can prevent a lot of problems in the future.”

The Yandex press service notes that in the legislative regulation of AI it is important “not to get ahead of reality” and respond to real risks, but not to phobias, relying on the practice of using technologies, which is only beginning to accumulate. “The main questions now being asked of AI concern ethics and the responsible use of technology by both developers and users. Business is aware of this and proposes self-regulatory measures: it initiates codes, declarations and industry standards. At the same time, technologies continue to develop in a favorable environment, and users receive the most useful services,” the company believes.

“Russia has already created a unique and workable mechanism for interaction between the state and the market through the national code of ethics for AI. It pilots approaches to future regulation. A good example is interaction on the issue of regulation of recommendation services, when the provisions of the law were supplemented by the ethics of the work of such services,” noted Andrey Neznamov.

Robin Product Development Director at SL Soft, Ivan Melnikov, considers initiatives to regulate artificial intelligence in the Russian Federation relevant. “The use of AI is associated with a large number of risks, ranging from possible “hallucinations”, inaccuracies in answers, and ending with ethical issues. Therefore, regulatory initiatives are timely. The introduction of regulatory measures should be aimed at reducing the number of cases of unfair use and abuse of AI technologies. This will increase the trust of citizens and enterprises and simplify the use of AI in important industries, such as medicine or data processing,” believes Ivan Melnikov.

Nikolay Afanasyev