Rosneft has stated that it is impossible to create "sovereign" artificial intelligence
Russia’s largest oil company, Rosneft, has stated that creating so-called “sovereign” or “national” artificial intelligence models without access to global technologies is virtually impossible.
This is stated in the company’s response to a draft law by Russia’s Ministry of Digital Development regarding the regulation of the artificial intelligence sector.
Rosneft emphasized that the requirements to develop and train neural networks exclusively by Russian companies and based on domestic data are technically unrealistic.
Among the main reasons:
- the lack of necessary computing infrastructure
- a shortage of large datasets of high-quality data in Russian
- dependence on foreign technological components
In light of this, the company proposed allowing the use of open data from the internet to train artificial intelligence models, regardless of where the servers are physically located.
Specifically, this includes the use of materials from open online encyclopedias.
Other industry organizations also supported this position. Business representatives note that currently, there are no artificial intelligence models in Russia that fully meet the criteria of “sovereignty.”
In addition, experts point to the ambiguity of the concept itself: the difference between “sovereign” and “national” artificial intelligence is not clearly defined, and the draft law lacks specific technical requirements and standards.
As a reminder, in Russia, the popularity of Asian messaging apps has surged following restrictions on access to Telegram.
