Standardisation of AI solutions: Russian–Indian frameworks for BRICS

The market for AI solutions in BRICS countries is showing one of the fastest growth rates in the world. However, regulatory differences and a lack of mutual recognition for solutions remain the biggest obstacles to transferring technology. Russia and India were the first BRICS nations to take action, launching a unified system for evaluating AI technologies. The foundation for this was the 2023 memorandum launching a "visa-free" regime for high-tech products developed by the two countries.

Expected financial impact of AI adoption on BRICS+ economies by 2030, USD trillion

Source: Yakov and Partners analysis

New trend: independent AI standards

Russia participates is the International Standardisation framework for AI. However, the provisions of this framework largely translate the Western regulatory approach, which is not always suitable for the needs of developing economies. Regulatory frameworks like the EU AI Act are based on specific ethical guidelines and may be at odds with the needs of BRICS countries.

The problem of universality

• Ethical gap In India, the priority is inclusivity and the accessibility of technology for even the poorest segments of society. Russia places an emphasis on technological sovereignty and security. Blindly copying the Western template can slow down the development of useful social technologies.
• Market barriers Global standards are used as a tool of soft power, with dependence on Western certification, leaving businesses in a vulnerable position.

Russia and India are creating an alternative centre of expertise, where the priority criteria are application efficiency and local context.

Joint white paper

Russia and India signed a Memorandum of Cooperation on 30 September 2023, which became the launchpad for synchronisation of regulatory frameworks for AI. In Russia, work is being driven by the Technical Committee for Standardisation of Artificial Intelligence (TC 164). The committee comprises over 70 organisations, including Sber, Yandex, and Vision Labs, as well as leading universities, research institutes, and the relevant authorities. The committee works as a mirror of the international subcommittee ISO/IEC JTC 1/SC 42, adapting global best practices in line with national interests.

The following sub-committees (SC) and working groups (WG) collaborate with India:

• SC 01 AI in Healthcare: standards for healthcare
• SC 03 AI in Transport and WG 04 Applied AI Technologies: agrotech and drones
• WG 03 AI System Quality: trust and security metrics

The Chamber for Indo Russo Technology Collaboration (CIRTC) is the Indian partner organisation involved in this cooperation.

Joint work is underway on White Papers, creating a roadmap for common approaches to evaluate AI solutions. The goal is to create a mirror system of standards to enable companies from one side to enter the market of the other without the need for dual licensing.

To achieve this, experts from Russia and India conduct market readiness audits. What does the White Paper include?

1. Catalogue of solutions Real industry AI applications created by Russian and Indian developers are analysed. This is a signal for businesses that the market is established and "national champions" have emerged.
2. A code of ethics "Red lines" in AI—boundaries that define acceptable and unacceptable areas for development or application—have been agreed upon, based on the Russian Code of Ethics for AI and India’s National AI Strategy. The principle of human-centric and explainable AI solutions has been approved, while "black boxes" will not be granted certification in key industries.
3. Technical foundation Russian GOST State Standards (including GOST R 59276-2020 Artificial Intelligence Systems, GOST R 59921.0-2022 Artificial Intelligence Systems in Clinical Medicine, and GOST R 59920-2021 Artificial Intelligence Systems in Agriculture) are proposed as a foundation. International and regional documents on standardisation in AI are also taken into account, including SC42, ITU, IEEE, etc.

Pilot industries: healthcare and agriculture

Healthcare and agriculture have been selected for the development of common AI standards: huge markets in which the cost of an algorithm error is critically high and the need for automation is growing exponentially. The successful harmonisation of solutions and approaches in these industries will simplify their scaling to other areas, such as fintech, manufacturing and smart cities.

Experts have established the key principle of comparison with a human operator: to obtain certification, AI must work at least as well as a qualified doctor or agronomist.

Healthcare

India and Russia both consider AI in healthcare as an important tool for social safety.

Key areas:

• clinical decision support system (CDSS) for diagnosis and patient routing
• imaging analysis, where Russian solutions for radiology are of interest to India
• telehealth and remote monitoring; AI serves as a filter that sends alarm signals to doctors in hard-to-reach regions of India

Practical cases:

• Russia SberMedAI, Cels (diagnostic assistant, medical imaging analysis)
• India Qure.ai (diagnostic solutions), Niramai (breast cancer screening)

Agriculture

India is an agricultural superpower, while Russia is a global leader in food exports. For the agriculture industry of both countries, AI provides a way to increase margins and reduce climate risks.

Key areas:

• analysis of satellite images for crop monitoring
• forecasting harvest yields based on Big Data
• autonomous control systems for agricultural machinery

Practical cases:

• Russia Rostselmash, Cognitive Agro Pilot (autopilot systems for tractors and combines), AssistAgo (agribusiness automation)
• India CropIn (smart farming), MyCrop (agribusiness automation)

The next stage is the creation of a network of accredited AI testing centres in Russia and India. Datasets for validation to established benchmarks—medical imaging libraries, satellite images, etc.—will serve as the basis for verification. If an algorithm displays efficiency "at least equal to that of a human", then it will be awarded the BRICS seal of approval.

New model of AI standards:

• Developer submits an AI solution to the country’s lab for assessment
• The lab carries out testing according to unified methods based on agreed datasets
• The certificate that is issued is automatically recognised by regulators in the partner country

What does this mean for business?

• Accelerated time to market Eliminating dual licensing reduces the time it takes to bring an AI solution to market by 6–12 months.
• Reduced R&D costs Unified testing methods and access to common datasets generates savings in development budgets.
• Seal of approval The BRICS Compliance Certificate is a clear measure of reliability for large B2B customers, counteracting the distrust of software developed by unknown actors.
• Scaling up to BRICS+ Solutions will gain simplified access to other countries of the alliance (Brazil, Egypt, Ethiopia, Iran, the United Arab Emirates, South Africa, and Saudi Arabia), a common market comprising around 45% of the world’s population.

Prospects: AI architecture for BRICS

The pilot project is a sandbox for which Russia and India are currently fine-tuning the mechanisms of interaction.

Roadmap for BRICS:

• Model for the alliance The bilateral model under development can be considered a ready-made template for technological integration for all BRICS participants.
• Common data market The next step is the formation of shared dataset libraries for training neural networks. This will make it possible to create sovereign fundamental models that are more relevant to the needs of BRICS states than their Western equivalents.
• Geographical expansion The architecture is initially designed as an open platform in which new BRICS members are already showing a keen interest.

Russia and India plan to leverage experience with joint laboratories to promote methods on a global level—comparable to international organisations such as the ISO, IEC, and ITU—offering an alternative to Western standards. In fact, a new technology macroregion is being established, where the rules of the game are determined by the domestic interests of alliance members.