By Hamza Faisal

Cybercriminals have found a new disguise for an old weapon: they are dressing up distributed denial-of-service (DDoS) bots as legitimate artificial intelligence crawlers, and slipping straight past the defences meant to stop them. Cybersecurity firm StormWall says it recorded the first major wave of such attacks in September and October 2026, in which malicious bots masqueraded under the names of official neural networks to crash company servers.

The findings, reported by TASS on October 31, mark the first time attacks in this form have hit major organisations in the Russian segment at scale. The twist is that the victims, in a sense, opened the door themselves — by welcoming AI bots in.

An open door meant for AI crawlers

The attack exploits a trade-off that thousands of websites have quietly made. Media outlets, online stores and other publishers have increasingly given free access to AI bots so that systems such as ChatGPT, DeepSeek and Gemini can read and index their pages when answering user questions. Being indexed by the big AI assistants has become a source of traffic, so blocking their crawlers is not an attractive option.

Attackers are now abusing that hospitality. By dressing their malicious bots up as supposedly useful, official neural-network algorithms, they present traffic that security systems assume is legitimate. Filters that would normally block a flood of junk requests wave the traffic through, and under millions of fake requests the targeted sites and apps are completely paralysed.

Media, shops and banks took the heaviest hits

StormWall's breakdown of the September and October attacks shows the campaign was broad rather than narrowly targeted. News portals and media platforms absorbed the largest share, at 32 percent, followed by online stores and major retail marketplaces at 26 percent. Banking and financial organisations accounted for 17 percent, telecommunications networks for 14 percent, and other online services for the remaining 11 percent.

The spread suggests a general-purpose weapon rather than an attack on a single sector. Media sites are a natural first target — they are enthusiastic adopters of AI indexing — but the same trick works against any organisation that has whitelisted AI crawlers.

The timing fits a wider pattern. StormWall's quarterly data showed AI-assisted attacks already made up 34 percent of all DDoS attacks worldwide in the second quarter of 2026, up from 22 percent the previous quarter, while overall DDoS attacks grew 108 percent year on year. The firm has separately reported that global DDoS attacks nearly tripled in 2025, from 6.6 million to 19.4 million incidents.

Verify the bot, not just the badge

StormWall's advice to website operators is blunt: do not take a bot's claimed identity at face value. Companies should implement deep verification systems that strictly check whether incoming IP addresses and requests genuinely belong to legitimate neural networks, rather than trusting a user-agent string or a familiar-sounding bot name.

In practice that means validating crawler traffic against the AI companies' own published IP ranges and verification methods, and treating anything that merely claims to be an AI assistant with the same suspicion as any other unverified visitor.

As AI assistants become a bigger part of how people discover information online, the incentive to impersonate them will only grow. The September and October attacks are a warning that the internet's newest gatekeepers — the bots that read the web for AI — have become a disguise worth stealing.