"Cyber villains will use AI systems for automatic hacking"
Russian IT companies, in conditions of cool relations with Europe and other once familiar partners, are able to switch to the markets of the Middle East, South America and Africa, "it is possible and necessary to work there," said Evgeny Kaspersky, co-founder and CEO of Kaspersky Lab, in an interview with Izvestia on the sidelines of the world's largest technology exhibition GITEX Global. He noted that Russian software remains the best in the world, but agreed that we lag behind competitors in creating hardware, i.e. physical components of computer systems. How does a businessman and programmer relate to artificial intelligence, what are the main dangers he sees in the use of AI by criminals and how he proposes to deal with it — in our interview.
"North America is bad, South America is good"
— You are positioned as an international company...
— No, an international company is the United Nations. International companies should be established by different states, such as Interpol. And we are a global company.
— Kaspersky Lab's roots are in Russia. As a co-founder and head of a global company, how do you assess our country's level of software development? Where are we in the world?
— I think you've heard the quote: "Russian programmers are the best in the world." If you do the right organization for development, testing, and other things, you can achieve great success.
— In the current conditions, what are the prospects for Russian developments in foreign markets?
— It depends on the region. North America is bad, South America is good. Europe is bad right now, but I think it's temporary. The Middle East is great. There are not very large markets in Africa, but you can and should work there. Asia depends on the country: there are countries where it is very good, and there are countries where it is difficult. Depending on how much they depend on a foreign or overseas partner.
— How do exhibitions like GITEX Global help promote our developments?
— GITEX is one of the leading IT exhibitions in the world, participation in which is absolutely mandatory. Previously, there was the CeBIT information technology exhibition, it was the world's largest IT exhibition a few years ago, and participation in it was also mandatory — everything was demonstrated compactly on one site. And GITEX is just as concentrated now. And we try to participate in all "Jitex" (GITEX Asia, GITEX Africa), because these are partners, these are clients. And what are we showing here? That's it!
"I really dislike the term 'artificial intelligence'
— Russian companies are now known precisely as software manufacturers. But there are significantly fewer offers for Russian equipment. Is this a problem?
— This is a huge problem. And her solution is beyond my capabilities. But I really hope that sooner or later this gap will be leveled. As far as I know, work is underway in this direction.
— A huge number of artificial intelligence projects are presented at the exhibition, including Russian companies. Although we know that you don't really like this phrase. What is your attitude to this trend, including the use of AI in development?
— Yes, I really dislike the term "artificial intelligence" because I advocate its strict definition. What is now called AI is complex algorithms, but it is not intelligence. Intelligence is the case when the system is ready to solve unpredictable tasks. And what we now call AI is just well-trained algorithms that solve a very narrow range of tasks. But this is something that has come into our lives and will remain forever, because these systems allow us to do so many things much more efficiently.
Let's say it's the control of industrial processes. Collecting telemetry processed by an artificial intelligence system allows you to see what the human eye cannot see. A person is not capable of this.
We have been developing such information protection systems for more than 20 years, that is, we started back when they were not called artificial intelligence. But even 20 years ago, it was already clear that the number of new malware on the Web was growing exponentially. We are currently catching about half a million new malware per day. And for this, we process approximately 15 million suspicious files. The human eye would never do that. How many millions of people do you need to sort through such a volume, if, for example, you sort through a hundred files a day?
As a result, we came up with the idea that we need to create an automatic malware recognition system. And we did it. It uses various filters, neural networks, recognizers, and so on.
Plus, we use an artificial intelligence system as a detector of anomalies, that is, behavior, in the Network. Of course, we also have all sorts of conversational assistants, assistants for IT security people. That is, we have a lot of different things that fall under the definition of artificial intelligence.
At the same time, cyber villains are using an artificial intelligence system to create new malicious products — the first frameworks have already appeared. In the future, artificial intelligence systems will probably be used for automatic hacking. Everything is already moving towards this.
— So these are the cyber threats of the future. And what others will there be?
— Rather, the question here is not what they will be, but how vicious they will be. Because everything has already happened. These will be attacks on infrastructure, critical infrastructure, disrupting its functionality. This is the most painful thing that can happen.
"There will be a massive introduction of cyberimmunity systems"
— Many people call you a visionary from the IT industry who was ahead of his time. What are your achievements over the past 20 years that you consider to be the most important?
— First of all, we transferred the development of our products to a new architecture, it is self-written, we invented it ourselves. As a result, the products have become much faster. It happened sometime in 2006. At the same time, we began to enter foreign markets much more actively.
Secondly, we guessed right and released a system on time to protect the industrial systems that are needed now.
The third achievement: the company has built absolutely transparent processes — telemetry is removed from everything: from processing, from sales, from finance. At any moment, we can find out how fast development is going, so we can ensure that products are released at the right time and of high quality, just by seeing what is happening. We have real-time sales statistics accurate to the country and to a specific product. That is, we can now click on the website, and I will find out how many home-made products have been sold, conditionally, in Uruguay today.
The fourth is global expansion. We started out as a small group and didn't speak English well. We were in a passive state — that is, we had partners and clients who came to us. But we started to slowly enter foreign markets, primarily Europe. We used our participation in the CeBIT exhibition for this purpose, and built a partner network more actively. After 2005, we set the goal of a global presence — to be everywhere, from Chile to Oceania. And we did it around 2012.
And finally, the fifth is a product that is still on the path to success, but I believe in it. This is the operating system we created.
— How will your concept of cyberimmunity develop and evolve (an approach to designing systems with mandatory protection against cyber attacks. — Izvestia)? And what will cybersecurity look like by 2035?
— Critical infrastructure, the Internet of things will start to be cyber-immune. It's already started, but it's still spot-on. And there will be a massive introduction of cyberimmunity systems.
Переведено сервисом «Яндекс Переводчик»
