ST. PETERSBURG, June 3. /TASS/. Companies already having advanced information search systems for the global network and developing the technologies and infrastructure necessary for this purpose will take a leading position in developing new AI systems in the coming years, Alexander Krainov, the Yandex Artificial Intelligence Technology Development Director, said at the St. Petersburg International Economic Forum (SPIEF).

"When you ask AI questions, in many cases the neural network goes into a search to obtain information that was not in the training data. At the same time, search engines are much more difficult to create than large language models, and there are currently only five global search engines in the world. For this reason, I think that Google is more likely to win the race in creating new AI systems than Anthropic or other companies that only develop neural networks," Krainov said.

Even the most advanced large language models will never replace search, the expert said. According to him, they are a kind of an add-on to search engines, an analogue of a web page where search results are displayed.

"Basically, AI is now fighting not with search engines, but with the results page. Therefore, in the future, there will always be a search engine under the 'hood' of such a system, which will continue to be an important systemic part of many business systems aimed at working with end consumers," Krainov added.