Business

Sber presents TV clip entirely created by its Kandinsky, GigaChat neural networks

Moscow, March 9 (IANS) Leading global financial institution Sber has released a TV clip which, in the first, has been entirely created by its own neural networks, Kandinsky and GigaChat.

Timed to coincide with International Women’s Day, the clip is a technological breakthrough for the Russian media and advertising industry, combining advanced artificial intelligence technologies to generate pristine visual imagery.

The Sber team has managed to implement a unique solution that was previously hard to imagine. The level of detail, light and composition all make it look as if the clip was produced by a team of artists and animators rather than by an algorithm. Some of the shots look like real film, but in fact, it was all produced by an AI.

It’s also important to note that Russian technology is not just in catch-up mode but is actually creating content which reflects the national culture.

The story is based on the idea that fairytale women with astonishing and outstanding qualities, such as Maria the Artful, Helen the Beautiful and Vasilisa the Wise, exist not only in books but actually surround us every day.

And AI technology is helping to demonstrate this in a vivid clip. The tune used in the clip was also composed by one of Sber’s generative models – its SymFormer neural network.

Russia is undoubtedly among the leaders in artificial intelligence today. It is helping to magnify human capabilities many times over. This project may set a precedent, proving that Russian AI technologies not only catch up with global trends but also create unique content reflecting national culture.

Sber managed to realise the mechanics, which became not just a tool but a real artist reflecting the Russian identity. The success of the company’s products in the domestic market shows, among other things, the potential of developers in the company who use AI to create television commercials.

–IANS

na/dpb

Related Articles

Back to top button

You cannot copy content of this page