TOKYO/SEOUL -- A series of major cyberattacks has hit businesses in Japan and South Korea in recent weeks, exposing millions of customer records and raising concerns that artificial intelligence could be helping hackers intensify their operations.

Targets have ranged from rental car services and major South Korean banks to Japanese retailers and restaurant chains. Experts are examining whether artificial intelligence is making attacks easier to carry out, though they have yet to establish clear links between the recent incidents or determine whether it played a direct role.

The hacks highlight a growing cyber threat facing Asian companies as data breaches and ransomware attacks become more frequent and potentially more damaging. The spread of generative AI has added to fears that attackers could automate parts of their operations and target victims on a larger scale.

Trend Micro, a Tokyo-listed cybersecurity firm, told Nikkei Asia it had compiled 600 cases of Japanese companies and organizations reporting unauthorized access and other cyber incidents by the end of September this year.

The figure, which records incidents by announcement date rather than when they occurred, had already surpassed last year's 593 and was approaching the record 644 cases logged in 2024. Incidents have been on a general upward trend since summer, with 73 cases reported in August and 86 in September.

On Tuesday, Japanese discount retailer Mr Max Holdings said personal data belonging to as many as 1.73 million customers may have been compromised after unauthorized access to a server used for its online shopping services.

Financial institutions have also been targeted. On Monday, Japan's Daiwa Securities said about 220,000 records had been leaked after hackers gained unauthorized access to a contractor last week. The compromised information included names, email addresses and account numbers from 110,000 customers.

CreditSights, a Fitch Solutions company, said seven financial institutions in South Korea -- including major lenders Shinhan Bank, KB Kookmin Bank and Hana Bank -- had been affected by security breaches, which it described as "a sector-wide cybersecurity incident." The research company said it could not yet assess the financial impact on the banks, but it predicted potential regulatory penalties and customer compensation payments, as well as increased spending on improving their online security.

Financial authorities in South Korea have flagged the risks presented by AI. "As AI continues to rapidly develop, new attack patterns and hacking techniques that exploit unforeseen vulnerabilities could spread quickly," the Financial Services Commission said in a statement on Monday.

President Lee Jae Myung on Tuesday also spoke about the growing threat of AI-driven attacks. "We have reached a scenario where even people with no special skills can use AI to easily carry out hacking attacks," he said.

Japan and South Korea were among the five most targeted markets in the Asia-Pacific region between June 2025 and June 2026, with 2,282 and 1,978 incidents, respectively, according to cybersecurity firm ThreatBook. Together, they accounted for about one-fifth of the attacks tracked by the company during the period.

But experts in Japan cautioned that evidence directly linking AI to individual attacks remained limited. "We are analyzing the causes and possible connections, but at this point, we cannot clearly say whether AI was involved," a Trend Micro representative said.

AI could be used in the preparatory stages of an attack, the representative added, but "finding evidence that it was used in the attack itself remained hard to trace." On Sunday, Nikkei Inc, owner of Nikkei Asia, said an employee's Microsoft 365 account had been compromised, allowing an attacker to send about 9,000 spoofed emails. The media company said names, email addresses and some email content were also believed to have been exposed.

One of the highest-profile cases in Japan involved Park24's car rental service Times Car, first announced on Sep. 25, where attackers gained access to information linked to 6.6 million accounts. About 1.6 million records were exposed, including identification documents such as driving licenses. Times Car had about 4.15 million members as of the end of August, with most believed to have been affected by the data breach.

The company had retained driving license information for seven years under its data-storage policy, even after customers had closed their accounts. The incident has raised concerns that the stolen material could be used for identity theft, including fraudulent credit card applications.

The breach has prompted calls for a class-action lawsuit over the company's handling of personal information, with more than 40,000 people joining a petition as of Monday.

The cluster of incidents reflected broader trends of Japanese companies becoming more vulnerable to cyberattacks and generative AI lowering barriers for overseas attackers, said Hirotake Nishine, an executive officer at Japanese cybersecurity company Josys. However, Nishine said it was not possible to determine commonalities as little information had been disclosed about how attackers gained access or the methods they used.

The Japanese language had previously offered some protection, Nishine said, as unnatural Japanese made phishing emails and impersonation attempts easier to detect.

"With generative AI, anyone can now easily produce Japanese that is indistinguishable from that of a native speaker," he said. "That barrier has effectively disappeared."