Russia and Belarus Forge AI Innovations Steeped in Cultural Heritage

Russia and Belarus are set to collaborate on the development of an artificial intelligence system that officials claim will be rooted in “traditional values,” as reported by the Belarusian state news agency BelTA on Friday.

Sergei Glazyev, the secretary of the supranational Union State, which includes both countries, stated, «Our objective is to create a sovereign AI system that is trustworthy, dependable, and capable of providing objective information.»

He cautioned that younger individuals are becoming increasingly “susceptible to manipulation” by foreign AI models, especially those from the U.S. and China.

Glazyev noted, “Recent extensive evaluations of Western-style AI exposed that one notable chatbot displayed racist and extremist inclinations, including the glorification of fascism,” although he did not cite specific instances or offer evidence for this claim.

According to Glazyev, the initiative aims to foster entrepreneurship and innovation while maintaining “fundamental and traditional values.” He did not elaborate on what he intended by «traditional values,» but he asserted that entrepreneurship lacking these values tends to be “destructive, irrational, and harmful.”

This announcement coincides with intensified efforts by Moscow and Minsk to establish technological and ideological autonomy from Western influences. Following the invasion of Ukraine and the subsequent imposition of extensive sanctions, Russia has particularly intensified its pursuit of digital sovereignty.

Meanwhile, independent researchers have raised concerns regarding the transparency and openness of AI systems developed in Russia. A recent study conducted by Ghent University in Belgium found that Russia’s leading language models, YandexGPT and GigaChat, exhibited the most significant levels of political censorship among 14 prominent AI models worldwide, including those from China.

The researchers noted that these Russian models frequently avoided politically sensitive inquiries, often stating an inability to discuss certain subjects or recommending that users seek information from other sources.