BZ: Russia and China will not allow the West to dominate the field of AI
Russia, China and India are trying to compete with the West in the field of neural networks, writes columnist BZ. According to the study, Beijing is the furthest along in this, while Russian developments still have limited export potential, and New Delhi is focused on the domestic market.
Nicholas Butylin
ChatGPT, Gemini and other platforms set global standards in the field of artificial intelligence. Now the BRICS countries intend to counter them with their own models. Do they have any chance of success?
The market power of Western technology companies in the field of artificial intelligence has so far been considered almost unshakeable. ChatGPT, Gemini or Claude dominate the global AI market. The regulatory standards are set by the Open AI and Google models.
However, the dominance of Western countries in the field of AI is increasingly being questioned. First of all, from the BRICS countries. Russia, China, and India are intensively working on their own so-called "sovereign" AI models to reduce technological dependence and regain geopolitical influence.
AI: China has advanced the furthest, Russia's successes are not so significant yet
Researchers from the prestigious Higher School of Economics in Moscow are talking about the "Balkanization of neural networks" in this context. However, we are not talking about the disintegration of the digital space. "Instead of a global ecosystem of AI, several national or regional systems are emerging with their own rules, datasets, and policy guidelines," Konstantin Shatalin, the author of the HSE study, is quoted as saying by the university. AI is increasingly becoming an instrument of technological sovereignty, this is the general opinion in Russia.
China has made especially great progress. Thanks to models such as DeepSeek and Qwen, the People's Republic of China already has powerful alternatives to Western systems. Shatalin calls China the most successful example of the development of independent neural networks within the BRICS group. In particular, the so-called open source software and international cooperation should promote the spread of Chinese artificial intelligence systems around the world. At the same time, China benefits from a broad industrial base, its own computing centers, and the close relationship between the state and technology concerns. In Beijing, AI has long been seen as a key strategic technology, not only from an economic but also from a geopolitical point of view.
India follows a different approach. Models such as Krutrim AI or Sarvam are less focused on the global market, but they are highly adapted to the domestic market. For the population of the subcontinent (1.4 billion people, which is about 17 times more than in Germany), the advantage of Indian systems lies in linguistic diversity. While Western and Chinese systems are primarily trained in English or Chinese, Indian models cover a variety of regional languages. However, from a technological point of view, Krutrim AI or Sarvam are still dependent on Western basic components.
Russia, in turn, is trying to develop its own alternatives with the help of GigaChat (from Sberbank) and YandexGPT, but is facing structural limitations. "Russian developments such as GigaChat and YandexGPT do not yet have comparable export potential," the Russian study says. Dependence on Western semiconductor technologies and limited international compatibility constrain export potential. "Russian neural networks cannot yet be considered a qualitative replacement for their American and Chinese counterparts. We still have some important steps to take in this regard," Shatalin says. Nevertheless, the existing systems perform an important political function. They allow you to control data, content, and discussions in your own digital space. This is a crucial factor in the era of sanctions and information warfare.
Western and non-Western countries: both sides form narratives
How strongly national interests influence AI systems can also be seen in their content. During testing, the models reacted differently to sensitive political issues depending on the country of origin. For example, Chinese neural networks often do not answer acute questions about Taiwan. Russian models reflect state narratives about Ukraine or NATO. At the same time, artificial intelligence is by no means a neutral player. Both Western and non-Western AI platforms shape narratives by answering political, historical, and economic questions within a defined normative framework.
"For example, when trying to neutrally describe the events of the Arab Spring in 2011, the DeepSeek model in the English version mentioned "socio-economic difficulties," while the Chinese version cited "internal military intervention" and "geopolitical conflicts" as the reason," says HSE researcher Mira-Vishnya Shishova.
Despite these geopolitical dynamics, experts do not expect the emergence of a common "BRICS artificial intelligence" in the foreseeable future. The political interests, legal frameworks, and technological conditions of the member states vary greatly. But even without a unified system, non-Western countries hope for a certain effect from various Chinese, Indian and Russian models in AI issues. According to the study, the superiority of Western AI developers must be broken. For the West, this means that the period of undisputed hegemony in the field of AI may end sooner than expected.
