MTS talked about using AI to create software code

By 2027, the share of software code created by neural networks in MTS systems will grow to 25%, now this figure is 8%. Pavel Voronin, the company's first vice president for technology, said this in an interview with Vedomosti newspaper.

According to him, "the share of AI-generated software code in the Russian market as a whole will increase at a similar pace." The popularity of these services, which is growing every year, will allow us to reach this level. By 2030, over 90% of programmers will use AI assistants to test and write code, Voronin predicts.

The experts interviewed by the publication also confirmed the trend towards increasing the use of AI to write lines of program code. In particular, as noted by Mikhail Khlebunov, Director of Servicepipe products, this will be relevant for applications where high performance is not needed. We are talking primarily about the admin interface, web portals and mobile applications.

Alexey Borshchov, head of AI products at Just AI, points out that, in fact, "AI is now taking over the functions of a june specialist, automating routine tasks and helping to write standard code blocks," which allows developers to "focus on more complex and creative aspects of work."

According to Kirill Pichtovnikov, Deputy General Director for Production and Technical Director of RTK IT, a complete replacement of IT specialists is unlikely to happen in the near future. AI may not always understand the nuances or specifics of a project, which increases the likelihood of error. AI, for example, can hallucinate or give incorrect recommendations to IT specialists. Therefore, security and ethical issues require human involvement, especially when it comes to making decisions affecting users' personal data, he concluded.