Yandex is the first in Russia has received the ISO/IEC 42001 certificate. He confirms that the company takes a responsible approach to creating artificial intelligence. An independent audit has established that the processes of developing and providing YandexGPT comply with all the requirements of the standard, including in the field of ethics and security.
YandexGPT is a family of large language models (LLM, Large Language Model) Yandex. They are used in more than 20 services and products of the company, including Search, AI assistant Alice, Neuroexpert, Advertising, Masterpiece, Smart Camera, AI assistants in the Market, Shop and Food. They are also used by Yandex Cloud clients on the Yandex Cloud AI Studio platform.
During the audit, experts made sure that Yandex properly ensures the security of YandexGPT user data, automatically monitors errors and anomalies in the operation of models, monitors the quality and ethics of responses, analyzes user feedback and maintains detailed documentation. The audit also confirmed that Yandex takes into account the ethical and social risks associated with the development of AI, and trains employees in the rules of responsible development.
Anna Zinchuk,
Head of the Information Security Compliance and Training Service
Yandex has always been very careful about everything related to artificial intelligence. We started building processes that would ensure the ethics and safety of AI long before the public announcement of YandexGPT. Getting a certificate is another confirmation that we are doing everything right. Yandex stands for responsible use of AI and is ready to share its experience with other Russian companies.
The principles that Yandex adheres to when developing and using AI are publicly available. In March 2024, Yandex, as part of the AI Alliance, developed and signed a declaration on responsible development and use of services based on generative AI. The company regularly introduces new measures to improve the security of its AI products. So, starting in the spring of 2025, Yandex attracts "white hackers" to check its generative neural networks and pays up to a million rubles for vulnerabilities found.
Please note that this press release is based on materials provided by the company. AK&M Information Agency shall not be held liable for its contents, nor for the legal and other consequences of its publication.
