The expert warned about the growth of cyber threats with the development of AI
- Новости
- Internet and technology
- The expert warned about the growth of cyber threats with the development of AI
The development of neural networks is changing not only business opportunities, but also the approaches of attackers to conducting cyber attacks. Artificial intelligence (AI) makes it possible to speed up vulnerability detection, system analysis, and attack preparation, so companies need to rethink data protection methods. Anton Dautov, an expert at the Relevant Internet agency, told "Izvestia" on September 16.
According to the expert, previously, many stages of cyber attacks required significant time and manual work by specialists. Now, with the help of AI, it is possible to analyze infrastructure faster, look for weaknesses, process large amounts of information and create auxiliary code.
The expert noted that the use of neural networks by attackers is already beyond the scope of theory. Modern models are used at different stages of attacks, from collecting information about the target to analyzing stolen data and preparing tools. At the same time, Dautov stressed, one should not assume that any major hacking is necessarily associated with the use of artificial intelligence. However, AI makes existing threats bigger and faster.
One vulnerable element can endanger the entire system.
According to the expert, recent incidents with large companies show that it is often enough for attackers to find one weak spot in the infrastructure. Such an entry point may be an outdated site component, an incorrectly configured API, an employee account, a password in the configuration file, a vulnerability in the container, or contractor access.
Dautov cited the situation with the travel company TEZ TOUR, which reported a hacker attack on its website on September 15, 2026. The company stated that after detecting suspicious activity, it isolated the ERP system and found no signs of compromising customer and partner data.
At the same time, the DataSuckers group announced a larger penetration and described the alleged chain of attack. However, according to the expert, this information has not been independently confirmed.
"This case shows the general principle of modern attacks: after gaining initial access, attackers begin to expand their capabilities inside the system," the expert explained.
Which threats are becoming the main ones
The expert noted that today the danger is not only hacking the site. Attackers can gain access to an employee's account and use it to upload data, find a vulnerability in the old API, discover secret keys in the code, or use contractor access. Therefore, the company's protection should be aimed not only at the external perimeter, but also at constant monitoring of processes within the infrastructure.
According to Dautov, special attention should be paid to the analysis of event logs, data transfer control, account protection, code verification and search for leaks of classified information.
It is also important to regularly check the access rights of employees and contractors, and monitor the settings of cloud services, containers, VPNs, and corporate storage.
How to safely use AI in a company
According to the expert, a separate area of protection is related to the use of neural networks themselves within organizations. When AI agents connect to CRM, corporate documents, mail, or other systems, new risks arise: data leaks, the model gaining redundant rights, executing unwanted commands, and attacks through special requests.
Dautov noted that artificial intelligence cannot be given unlimited access to corporate resources. For AI systems, the same security principles should apply as for employees: minimum necessary rights, control of actions, logging of operations and additional confirmation of critical commands.
Cybersecurity will become a confrontation between automated systems
According to the expert, the main change in recent years is not the emergence of a new type of attackers, but an increase in the speed and scale of attacks. AI allows you to find vulnerabilities faster and simultaneously work with a large number of targets. In response, companies will be forced to make greater use of automated analysis, monitoring systems, and threat forecasting tools.
In the coming years, cybersecurity will evolve as a confrontation between automated systems: some will look for weaknesses, while others will discover and close them before they are exploited by intruders.
Переведено сервисом «Яндекс Переводчик»