The 2025 Cybersecurity Report released by VSEC indicates that nearly half of all DDoS attacks in Vietnam over the past year involved artificial intelligence.
DDoS attacks aim to disrupt the normal traffic of targeted servers, services, or networks by overwhelming them with massive volumes of internet traffic. According to the 2025 Cybersecurity Report published by VSEC — a member of G Group — AI was involved in approximately 117,000 DDoS attacks in Vietnam during 2025.
AI is not only increasingly used in cyberattack campaigns but also in online fraud schemes targeting users. VSEC noted that in 2025, AI-assisted fraud cases in Vietnam and across the region increased dramatically.
Globally, more than 28 million cyberattacks involving AI were recorded in 2025, representing a 72% increase compared to the previous year. These attacks achieved a success rate of 70%, significantly higher than the 47.6% success rate of traditional manual attack methods.
According to Phan Hoang Giap, Deputy CEO and CTO of VSEC, AI is transforming the way cyberattacks are conducted. “Attackers use AI to increase speed and efficiency. If handled manually, organizations will struggle to keep up,” he said.
AI allows attackers to automate multiple stages of the attack chain, from information gathering and vulnerability exploitation to personalized phishing content and deeper system infiltration. Human involvement is mainly limited to fine-tuning and optimization.
This shift is making traditional defensive measures increasingly ineffective, while security teams at Security Operations Centers (SOC) are facing growing pressure.
Looking ahead to cybersecurity trends in 2026, the report predicts that AI will be increasingly used to strengthen defensive capabilities. The question is no longer whether to use AI, but where to apply it effectively.
The first priority is event detection and correlation. According to the report, applying AI and automation within SOC operations can reduce incident detection and response time by 33–43% compared to environments without AI.
AI can also classify alerts based on priority levels and recommend response actions based on contextual analysis, reducing the workload on security teams and improving average response time.
However, AI is not a “silver bullet.” Without standardized data, optimized operational processes, and skilled teams capable of managing AI systems, organizations may face increased costs without proportional benefits.
As AI becomes widely deployed across enterprises, the AI systems themselves may become targets of attack, including data poisoning, model theft, prompt injection, or abuse of AI agents to gain unauthorized access to system resources.
