AI-Driven Cyber Attacks Target Japan with Automation and Advanced Techniques
Cybersecurity researchers have detailed three cyber attacks targeting organizations in Japan, highlighting the significant role of artificial intelligence (AI) in these operations, either autonomously or as support for human operators.
TrendAI analyzed three exposed server environments used by threat actors for these malicious cyber operations, noting that these attacks differed from previous incidents that employed ARTEX, an autonomous AI penetration testing framework, to identify targets in South Korea and Japan.
According to security researchers Bakuei Matsukawa and Joseph C Chen, “These environments held reconnaissance and attack tools, scripts, target lists, results, logs, and command histories. They ranged from workspaces used largely by hand by skilled operators to AI-agent platforms that automated the entire workflow, including command and control.”
The first attack involved the exploitation of CyberStrikeAI, an open-source penetration testing tool previously used in attacks against Fortinet FortiGate appliances. This activity was concentrated in October 2026.
The AI-driven operation utilized CyberStrikeAI to automate various aspects of the attack lifecycle, including reconnaissance, vulnerability discovery, exploitation, command-and-control (C2), persistence, data exfiltration, and automated reporting.
This process involved delegating tasks among 44 agents to manage different aspects of the attack, including generating forms and fake login pages in Japanese.
Trend Micro’s enterprise security division reported that at least 19 identifiable organizations, including 16 in Japan, were subjected to validation activity, with each organization receiving a report in a standardized format, indicating a structured process rather than random notes from an operator.
The second case was characterized as a human-led attack, with assistance from a large language model (LLM) to expedite planning and tool development. The campaign targeted AI workflow platforms and inference services to acquire large amounts of inference capacity at low cost.
One intrusion targeted a geographic information system (GIS) server that was exposed with default credentials, leading to the theft of cadastral and commercial records, Windows credentials, and email data. This campaign lasted for 24 days in August and September 2026.
The third instance showed AI’s role limited to developing tools and scripts for the operator, indicating a more traditional approach to targeting edge devices from Cisco, Citrix, F5, and Yamaha using known vulnerabilities and credentials in configuration files.
The threat actor reportedly executed at least six campaigns from January to October 2026, targeting exposed VPN devices and WebLogic servers. Researchers noted that the operator was experienced, with established practices for automation and evidence management.
Currently, there is no evidence suggesting that these three incidents share a common operator. It remains unclear whether the attacks have affected other regions or which AI models were employed in these operations.
Despite this uncertainty, the attacks are not limited to Japan and South Korea. The operator behind the first case has also targeted organizations in Taiwan and Hong Kong, while the third case included organizations in the U.K., France, Italy, Qatar, and Colombia.
TrendAI stated, “AI has increased the attackers’ workforce, but the work still consists of looking for unlocked doors. The types of doors and the weaknesses that leave them open have not changed. What has changed is the ability to try thousands at once, adapt to the target country’s language, and continue without fatigue.”
This development coincides with the emergence of multiple campaigns where threat actors have integrated agentic AI into existing offensive security workflows, enhancing their capabilities in research, malware development, and social engineering.
AI is increasingly being utilized to accelerate reconnaissance, automate decision-making, and industrialize social engineering, significantly reducing the time from vulnerability discovery to exploitation. Microsoft’s 2026 Digital Defense Report noted that AI has decreased the median time from vulnerability discovery to weaponization to below 24 hours.
The adoption of AI by malicious actors is diminishing the time, expertise, and cost required to identify and exploit vulnerabilities, allowing both attackers and defenders to operate with greater speed and scale.
Trend Micro emphasized that the key issue is not merely whether an actor has access to advanced AI but how effectively those capabilities can be sustained and utilized. They stated, “AI is no longer an experimental capability for APT groups; it is an operational component embedded directly into campaigns.”
The rising demand for AI tools among threat actors has also created new monetization opportunities, including the sale of stolen AI provider session tokens and offensive AI tools, further commoditizing cybercrime.
Recently, Microsoft reported that a China-state-aligned threat actor compromised a ConnectWise cloud backup application, using ARTEX to access over 2,250 Microsoft 365 tenants. This activity, attributed to Storm-0940, illustrates how a threat actor can leverage AI tools to quickly attempt to compromise numerous organizations.
Local media reports from South Korea indicate that authorities have identified seven suspects linked to a cybercrime operation that breached over 90 organizations, primarily involving ARTEX. The alleged mastermind, a 32-year-old, previously worked at a cybersecurity company, gaining expertise in penetration testing.
In 38 incidents, the threat actors demanded ransom payments of 360 million won ($270,000) to avoid leaking stolen data, with over 8.2 million records believed to have been compromised.
AI-oriented campaigns documented in recent weeks include various attacks employing advanced techniques and tools, demonstrating the evolving landscape of cyber threats.
Source: The Hacker News