BRICS should develop own language models

The focus of BRICS cooperation should not only be on developing new technologies, but also on sharing real educational experience, said Aleksei Semenov, a professor at Lomonosov Moscow State University. Semenov is also an academician of the Russian Academy of Sciences, and an academician of the Russian Academy of Education.

Currently, the majority of artificial intelligence research is concentrated in the United States, with some attempts in Europe, but BRICS countries, including Russia, have seen comparatively less progress in this area, Semenov noted. One critical issue, he pointed out, is data bias. When AI is trained using data primarily from English-speaking countries, it can lead to biased outcomes. He explained that this problem is amplified on the internet, where biased data, such as fake news, spreads quickly and influences other information.

To counter this, BRICS countries should leverage open resources and open-source technologies to develop AI models using national languages, the professor suggested. This would help reduce the bias inherent in relying solely on English-language data. Additionally, he emphasized, it’s important to focus on the local education system in BRICS countries, sharing knowledge and experience from teachers, students, and communities. Semenov highlighted that China, in particular, has significant potential in this regard and could play a leading role in this exchange of educational expertise.