A new and, according to experts, extremely dangerous trend has emerged in the cybersecurity sphere: cybercriminals have begun using bots disguised as artificial intelligence agents to knock company servers out of service. Analysts at the international information security company StormWall recorded the first major DDoS attacks disguised as official neural networks in September and October 2026. According to TASS reports, although similar schemes had previously been observed in a number of countries worldwide, they were applied to major organizations in this form for the first time.

Who Became the Target: Media, Marketplaces, and Banks

Cyberattacks primarily affected four key sectors of the digital economy. The main blow fell on the mass media, accounting for 32 percent of all recorded attacks, primarily targeting news portals and media platforms. Online stores and large retail marketplaces took second place (26 percent), banking and financial organizations including payment systems and financial services ranked third (17 percent), and the telecommunications sector — telecom operators and internet service providers — closed out the list (14 percent).

How the Scheme Works: Why Protection Fails

The most dangerous aspect of this attack is that companies unknowingly “open the door” to cybercriminals themselves. Today, many media outlets and online stores intentionally provide free access to their web resources for artificial intelligence bots so that systems like ChatGPT, DeepSeek, and Gemini can index their content when generating user responses. Hackers exploit this very opportunity: they disguise malicious bots as supposedly useful and official neural network algorithms. Security systems, mistaking these requests for legitimate artificial intelligence, do not block them, and under the onslaught of millions of fake requests, the operation of websites and applications is completely paralyzed.

Expert Recommendations: Deep Verification

Experts warn that even if companies are interested in having their resources indexed by artificial intelligence, they are obligated to implement deep verification systems. Such mechanisms must strictly check whether incoming IP addresses and requests truly belong to genuine neural networks rather than malicious bots disguised as them. Without such verification, the desire to be “visible” to AI turns into a vulnerability that attackers can exploit to massively bring down infrastructure.