GMO Internet Group’s security arm said in a 19 October blog post that Artex, a China-linked AI penetration-testing tool, had been used in attacks on Japanese companies, The Straits Times reported. The finding puts Japan alongside South Korea, where a separate investigation has implicated Artex in attacks on financial institutions since late September.
Key points
- GMO says Artex was used in recent attacks on Japanese companies.
- Its researchers found six servers launching AI-powered attacks aimed almost exclusively at Japan.
- Eight Japanese firms disclosed leaked or possibly leaked data in September and October.
GMO identifies six attack servers
Daiki Fukumori, director of the cybercrime countermeasures center at GMO Cybersecurity by Ierae Inc., said researchers had confirmed Artex’s use in many recent Japanese attacks, Bloomberg reported on 20 October. The company investigated the attacks with external organisations, The Economic Times reported.
Fukumori said GMO found six internet servers launching AI-powered attacks. He described the activity as “almost exclusively targeting Japan” and broadly grouped the servers by the industries they attacked, including finance, e-commerce and logistics.
About 80 Japanese companies in insurance, finance and logistics were attacked, Chosunbiz reported. It said an investigation commissioned by affected companies found traces of Artex on all six overseas public servers used in attacks between 18 September and 18 October.
Eight firms disclosed leaked or possibly leaked data in September and October, The Straits Times reported. The attacks targeted companies in retail, logistics and finance. Separately, Japanese businesses spanning food-service groups and a vehicle-rental firm have reported illicit system entry or data breaches in recent weeks, The Economic Times reported.
Artex’s use in South Korea
Artex was built for AI-assisted security testing intended to find weaknesses in computer systems, its developer has said. Chosunbiz described it as a Chinese-language autonomous tool that can analyse targets, plan intrusion routes and run security tools when connected to a large language model.
In South Korea, security firm CrowdStrike assessed that attackers used Artex alongside a large language model in attacks on financial institutions, Chosunbiz reported. AI Affairs previously reported Lee’s order for a probe into the bank hacks amid questions about AI use.
Artex’s developer subsequently stopped programmers from accessing its source code and said it would no longer update the project, The Economic Times reported. The developer said using the tool for attacks was “entirely contrary to” its intentions. AI Affairs reported its decision to take Artex closed-source after the Korean bank hacks.
Furukawa examines malicious AI use
Toshiharu Furukawa, Japan’s minister in charge of cybersecurity, said at a 20 October press conference that the government was examining whether AI had been used maliciously after hearing from affected companies, The Straits Times reported. He pledged to continue efforts to prevent cyberattacks.
