Chinese AI tool linked to cyberattacks

A Chinese AI tool named Artex has been linked to a series of cyberattacks targeting Japanese companies, according to the Tokyo-based subsidiary of a major internet group. The attacks have affected dozens of businesses, including restaurant chains and a car rental operator, with reports of unauthorized access and potential data breaches involving millions of customers.

Cybersecurity firm confirms AI tool's role

The cybersecurity branch of GMO Internet Group confirmed the use of Artex after investigating the attacks in collaboration with external organizations. Daiki Fukumori of GMO wrote on the company's blog that the attacks appear to have targeted almost exclusively Japan, suggesting a focused effort against Japanese businesses.

Artex designed for security testing

Artex was designed as an AI tool to test cybersecurity defenses and help organizations identify potential weaknesses. However, the Chinese developers recently stopped allowing programmers to access its source code, citing abusive use by "malicious actors" and stating that the tool would no longer be updated or maintained.

Previous links to cyberattacks in South Korea

South Korea had previously linked Artex to several cyberattacks on financial institutions. The developers of Artex stated that its use for cyberattacks is "totally contrary" to their intentions and emphasized that the project will transition to closed-source code to prevent further misuse.

Six servers identified as attack sources

Six servers were identified as the source of the attacks, according to Daiki Fukumori. GMO's subsidiary categorized these servers by sector, such as finance, e-commerce, and logistics, but did not specify their locations or the identities of those exploiting them. The company shared the information with authorities to help prevent further damage.

Japanese government urges stronger defenses

The Japanese government has called on companies to strengthen their cybersecurity defenses in response to the attacks. This follows growing concerns about the potential misuse of AI tools designed for security testing, highlighting the need for better safeguards against such threats.