Non-standard application: AI from the Russian Federation has received an ISO quality certificate for the first time

Artificial intelligence from Russia has received the international ISO quality certificate for the first time. So far, it has only been issued to companies from unfriendly countries. Yandex became the first in Russia, the AI market told Izvestia. This certificate shows that artificial intelligence development is conducted ethically, user data is not leaked, and errors and anomalies in AI are being monitored, experts explain. They do not exclude that such certification will be of interest to other market participants.

What does an AI Development Quality certificate mean?

For the first time, a Russian company has received a certificate of compliance with the international ISO standard, which sets requirements for the quality of development and security of artificial intelligence. Yandex told Izvestia about this.

The representative office of TUV AUSTRIA Standards & Compliance confirmed this information. They noted that the company has passed an independent audit and certification according to the international standard ISO/IEC 42001 "Information Technology — Artificial Intelligence — Management system".

Photo: IZVESTIA/Pavel Volkov

ISO/IEC 42001 is an international standard for management systems in organizations that develop or use artificial intelligence—based solutions. He evaluates not the AI products themselves, but the approach to their creation and implementation: how transparently and safely the company operates, how effective its processes are. In particular, the standard describes how user data security is ensured, how errors and anomalies in AI are monitored, the level of quality control and ethics of the language model's responses, whether developers take into account the ethical and social risks associated with the development of artificial intelligence, and much more.

Market participants interviewed by Izvestia note that according to the standard introduced in 2023, only a few companies and services have been certified so far, and all of them are from unfriendly countries. Among them are Microsoft, Google Cloud, Amazon Web Services, Samsung SDS. TUV AUSTRIA Standards & Compliance said that no more than 50 such international certificates have been issued in the world.

Photo: IZVESTIA/Dmitry Korotaev

Yandex began building processes that would ensure the ethics and security of AI long before the public announcement of the generative neural network YandexGPT, said Anna Zinchuk, head of the company's compliance and information security training service. The organization advocates responsible use of AI and is ready to share its experience with other Russian companies, she noted.

Izvestia sent requests to the Ministry of Finance, Sber, which is actively developing technologies based on artificial intelligence, as well as to the largest IT and telecom companies.

Playing by the rules: Russia has developed a code of ethics for working with AI for financial organizations

The document is designed to ensure the safety of technology application in the market

What does AI certification give to the end user?

The international standard establishes a concept and clear framework for responsible management of AI technologies, TUV AUSTRIA Standards & Compliance told Izvestia. The certificate confirms the company's commitment to the safety and reliability of its AI solutions. It is issued in case of positive results of an independent audit, which should confirm that the organization is taking all reasonable management measures consistent with the best international practices in order to manage risks and opportunities, they noted.

The attitude of market participants towards ISO certification of AI is ambiguous, some consider it only an image element and do not see any practical need for it.

Photo: IZVESTIA/Anna Selina

— At the moment, we have not passed the ISO/IEC 42001 certification, but we are actively working on a number of internal initiatives that correspond to its key aspects. We have a special map (AI implementation risk heat map) that allows us to systematize and minimize risks at all stages of AI implementation," Roman Khazeev, Director of Digital Technologies at Rostelecom, told Izvestia.

However, experts believe that in the future, more and more participants in the Russian digital market will show interest in certification.

— Such certificates are important, for example, from the point of view of investors: they provide an additional guarantee that something will not happen to AI that will collapse capitalization. Compliance with international standards is also interesting for potential partners and customers, in particular from friendly countries. Finally, this is also important for ordinary users, who are now often wary of AI products, not being sure of their safety," says Denis Kuskov, CEO of TelecomDaily.

Photo: IZVESTIA/Sergey Lantyukhov

In artificial intelligence, information security is not expressed in the usual format — it concerns the ethics of its behavior, said Anton Averyanov, CEO of the ST IT Group of companies, TechNet NTI market expert.

— Large language models contain a lot of data on which they were trained. And if there was any unethical or illegal information in them, then the user can get it," he added.

Scoring the slot machine: how students mastered the course with the help of an AI professor

Their results were higher than those who had been trained by an ordinary teacher.

Tatiana Mosendz, Director of Digitalization at the Russian School of Management, in turn, recalled that domestic standards for AI already exist. In January 2025, a national standard was introduced in Russia, the text of which is identical to ISO/IEC 42001. GOST applies to all organizations developing or using AI systems, and defines similar principles of management, ethics and responsibility.

"The appearance of a certificate of this level motivates companies to systematically build technological and production processes, identifies risk areas and development," she believes.

For end users, ISO/IEC 42001 certification means increased trust, security, and quality of interaction with AI products, says Yaroslav Seliverstov, a leading AI specialist at University 2035.

Photo: TASS/Annette Riedl

— First of all, this confirms that user data is protected at the level of international requirements. Automatic monitoring of errors and anomalies reduces the risk of receiving malicious or incorrect responses. Feedback systems make it possible to quickly correct AI behavioral errors and take into account real user feedback. The principles of ethics enshrined in the company ensure that AI does not issue offensive, discriminatory or misinforming responses," he added.

According to him, this is especially important when using AI in educational, medical, legal and advertising products. For families with children, it also reduces the likelihood of accidental access to unethical content, the expert concluded.