Hackers are using AI to significantly enhance their attacks, resulting in overwhelming, record-breaking distributed denial of service (DDOS) attacks.
This is according to Nuno Ceitil, consulting systems engineer at Netscout, who said hackers are gaining the upper hand by using AI to automate and improve their attacks.
Speaking at the ITWeb Security Summit yesterday, he said: “People use AI virtually every day, and attackers are doing the same to maximise the impact of their attacks. Attack groups have their own LLMs, with no guardrails and skewed moral and ethical boundaries.”
He cited the Aisuru botnet – one of the fastest and most disruptive IOT botnets in history, which generated a series of multi-terabit DDOS attacks in 2025 and early 2026.
“Aisuru caught a lot of people off guard,” Ceitil said. “It holds the record for its 31.4Tbps attack, peaking at 14.1 billion packets per second.”
He added that enhanced DDOS for hire now offers full attack customisation and uses conversational AI to replace the manual heavy lifting that used to be required to launch attacks.
Attack groups have their own LLMs, with no guardrails and skewed moral and ethical boundaries.
Nuno Ceitil, Netscout
Carpet bombing has also become a common pattern, he said. In the past, attackers might launch a single 100Gbps attack against a single host, but this would be quickly detected and mitigated. In a carpet bombing attack, a 100Gbps attack directed at 1 000 hosts simultaneously would avoid almost all bandwidth thresholds in DDOS detection systems but still cause the same amount of traffic on the network.
“Unfortunately, the bad guys may always be ahead,” Ceitil said.
Netscout advocates a multilayered hybrid defence strategy to combat increasingly frequent and complex DDOS attacks.
Share
