Kaspersky Zheng Qiliang: The non-standard use of AI may bring huge risks to human beings

On June 9, it was reported that Kaspersky held a media conference in Beijing on August 8, sharing the change trend of its possible threat pattern to the security industry in 2023, and explaining and looking forward to Kaspersky's business development in recent years.Zheng Qiliang, General Manager of Kaspersky Greater China20% Zheng Qiliang, General Manager of Kaspersky Greater ChinaSince 2013, Kaspersky's business in enterprise level market (B2B) has developed rapidly, and it has become the main business with a proportion of more than 80%

On June 9, it was reported that Kaspersky held a media conference in Beijing on August 8, sharing the change trend of its possible threat pattern to the security industry in 2023, and explaining and looking forward to Kaspersky's business development in recent years.

Zheng Qiliang, General Manager of Kaspersky Greater China

20% Zheng Qiliang, General Manager of Kaspersky Greater China

Since 2013, Kaspersky's business in enterprise level market (B2B) has developed rapidly, and it has become the main business with a proportion of more than 80%. In the overall business of Kaspersky in 2022, more than 50% of the business will come from the major customer market, and the other 30% will come from the SMB business. Because the domestic stand-alone market is free, the B2C business, which was once the largest in the early days, now only accounts for a small part.

"In recent years, Kaspersky has maintained an annual business growth rate of 20% -30%, of which 60% -70% comes from the contribution of the major customer market," he said.

Supporting ICT is Kaspersky's important mission

"At each market stage, we have to face different challenges." Zheng Qiliang said that Kaspersky is currently facing a biggest challenge and also a biggest opportunity, that is, the historical moment of promoting ICT in China. In addition to doing a good job in the existing business, we should also strive to enter the circle of ICT.

"Kaspersky fully understands and supports the government's demands, strives to actively pass the product and security tests of domestic Sinotron, and then vigorously supports domestic users. This is Kaspersky's most important mission in China at present," said Zheng Qiliang.

Referring to the market target in 2023, Zheng Qiliang said that he hoped that Kaspersky Greater China would strive to achieve more than 50% growth in OEM business, more than 35% growth in key customer market and 10% -15% growth in SMB market through efforts. "Although the B2C standalone business is relatively small at present, Kaspersky still maintains an annual growth performance of 10% -15%," he said.

A more complex security landscape may emerge in 2023

Kaspersky Senior Security Researcher Jin Ye introduced Kaspersky's outlook on Advanced Persistent Threats (APT) and defined the changes in the threat landscape that will occur in 2023. In the forecast for the next year, it includes attacks against satellite technology, email servers, an increase in destructive attacks and data leaks, hacker attacks against drones, and the next large-scale cyber espionage attack.

Jin Ye, Kaspersky Senior Safety Researcher

According to statistics, large-scale and significant online epidemics usually occur every 6 to 7 years. The last such event was the infamous WannaCry ransomware worm, which utilized the extremely powerful EternalBlue vulnerability to automatically propagate to vulnerable computers. Kaspersky researchers believe that the next event similar to "WannaCry" is highly likely to occur near 2023.

"Obviously, in 2022, the world geopolitical order has undergone major changes and ushered in a new era of instability. Some of our predictions focus on how this instability will turn into evil network activities, while other predictions reflect our views on which new Attack vector attackers will explore. Better preparation means better flexibility. We hope that our assessment of the future will enable defenders to increase Strengthen their systems to more effectively resist network attacks Jin Ye said so.

The non-standard use of AI may bring huge risks to humanity

With the increasing popularity of ChatGPT. At present, almost all Almost all are using and testing the AI big model related to ChatGPT in various ways. During this process, it may pose certain security risks.

Firstly, the most important risk may be information leakage. Many enterprises and individuals are trying to input some of their data and information onto the internet, ignoring the information leakage caused by this action. Fortunately, many professional enterprises have begun to block internet services in this area to avoid the risk of information leakage.

Secondly, if humans directly use the answers provided by AI models such as ChatGPT as a decision-making basis, it may also lead to significant risks. At present, in many tests, ChatGPT can also make mistakes in some very simple computational steps, and even fabricate some reference materials. Once humans abuse such results, it may bring potential risks to themselves.

However, Zheng Qiliang also stated that everything has one side and two sides, and if properly controlled, AI is indeed a good tool. Kaspersky has started to use AI technology to help him complete the defense and analysis of network security in the past decade, and its application is very mature.

At present, more than 90% of attacks in Kaspersky system are handled by AI means.

In specific general business, Kaspersky's engineers also tried to ask ChatGPT to help generate reports for virus detection, which really saved a lot of labor costs.

At the same time, in the analysis of threat attacks, a large type of work is to compare data, compare what kind of organization has done what kind of behavior, these tasks may allow AI to effectively share a portion.

In Zheng Qiliang's view, Kaspersky will not only pay attention to the development of the entire AI industry and use AI applications to help complete more automation work, but also pay attention to whether AI will bring some unpredictable threats.

ChatGPT is a good thing, but humans must make good use of it or use it carefully. Humans need to learn how to defeat magic with magic, "Jin Ye said.

It is important to put AI in the right direction. With proper management, it can indeed help us optimize workflow and improve work efficiency. But once AI loses control, it is another world. We all don't want to see human conflicts coming from AI's misjudgment, which is completely possible. There are already many scholars and prophets discussing this matter, "Zheng Qiliang said.

Zheng Qiliang believes that "if inappropriate use of AI is found to have some 'specific adverse purposes', it may be necessary to judge its abnormal behavior like treating drones and make measures. Ultimately, we hope that the correct use of AI can improve our human life