85540

Japan and U.S. to start joint AI cyberattack research

January 03,2025

By:KENSUKE WATANABE

TOKYO -- Japan and the U.S. will partner on research into cyberattacks that exploit artificial intelligence, Nikkei has learned, looking to combine advanced American technology with Tokyo's data on incidents outside English-speaking countries.

Japan's National Institute of Information and Communications Technology, which is overseen by the Ministry of Internal Affairs and Communications, will establish a research facility in Washington. This work is expected to begin as early as the fiscal year starting in April.

Researchers will be sent from NICT to cooperate with Mitre, a nonprofit organization that conducts research and development with U.S. government funding.

U.S. cyberdefense research focuses on attacks in English-speaking nations, and it lacks data on incidents beyond that sphere. But improvements in translation technology using generative AI have heightened the risk that methods used outside English-speaking countries will be directed at the U.S. Japan will contribute its relevant data to the research.

Washington regards cyberattacks that take advantage of AI as having a potential impact similar to a national security risk. The U.S. looks to expand joint research with an eye on China, positioning Japan as a core partner in dealing with cyberattacks in Asia.

The U.S. has far more specialized personnel in this field than does Japan, according to Tatsuya Mori, a Waseda University professor who is knowledgeable about security technology.

"There are few research opportunities in Japan for cyberattacks that exploit AI, so this initiative should be fruitful," Mori said.

Technological innovations sparked by generative AI have made it easier for individuals to create malware and sophisticated email scams. A global average of cyberattacks on companies in July to September 2024 was up 75% year-on-year, Israeli security giant Check Point Software Technologies reports.

Cybercrimes in Japan also are rising. Japan-based security company Trend Micro, based on a September report from the National Police Agency, noted that generative AI is improving cyberattack capabilities.

AI could be used to coordinate and optimize distributed denial-of-service attacks, which flood servers with traffic to take them down, or conduct ransomware attacks by finding software vulnerabilities. Using deepfakes to impersonate other people and steal confidential information is another concern.

Fewer than 20% of Japanese businesses have plans for active use of generative AI, lower than the 46% in the U.S. and 71% in China, a survey by the internal affairs ministry found. But 70% of surveyed companies in Japan said using generative AI will increase security risks, compared with nearly 80% in the U.S. and China.

Tokyo, which considers it necessary to promote the use of generative AI while reducing risks, plans to create guidelines for countermeasures against AI cyberattacks as early as fiscal 2025. The government also will work on devising a system that uses AI to collect information on cyberattacks worldwide and detect signs of attacks targeting Japan as early as possible.