MOSCOW, January 26. /TASS/. A direct confrontation between two AIs – one on the hacker side and the other on the cyber defense side – is likely in the future, Alexey Novikov, Managing Director at Russia’s information security company Positive Technologies, said in an interview with TASS.

He acknowledged that the use of artificial intelligence in cyber defense is still lagging, which is particularly noticeable in large corporations due to lengthy implementation cycles for various solutions. Many customers, he added, still counter hackers' AI-driven attacks "in the old-fashioned way." "That is, by using traditional tools and manual processes that were not originally designed for such a high level of automation on the part of attackers," Novikov explained.

Citing the example of AI-driven voice schemes in the fake boss scam, which has been known since late 2023, he noted that hackers, "unfortunately, are quicker and more active in adopting modern, widely available tools, with AI already having entered everyday use."

"We will see a confrontation between artificial intelligence systems on both sides of the barricades – the defensive side and the offensive side. This will be a competition not so much between individual tools, but between the quality of data, sophistication of mathematical models, and how deeply such systems are embedded in real-world security processes," Novikov said when asked about expectations for 2026 and the longer term.

Security specialists began using AI to search for vulnerabilities in 2025, even though this remains an experimental field, he noted. "However, it is precisely these experiments that usually signal which direction an attack will take. It is logical to expect that attackers will soon begin to widely employ similar approaches – for automated searches for vulnerabilities, configuration errors, and architectural weaknesses. It is not a question of if, but when: it may take months or a year, but the trend is already clear," the specialist concluded.