A 15-year-old boy in Japan has been arrested for using artificial intelligence to help him build a programme that forcibly deleted tens of thousands of accounts on an anime streaming service.
Police described the incident as a deliberate and persistent cyberattack that forced a month-long service suspension.
The high school student from Tokorozawa, near Tokyo, was accused of fraudulent obstruction of business after sending false commands to the servers of Bandai Namco Filmworks, a subsidiary of Japan's largest toymaker Bandai Namco Holdings, on 4 November 2025.
The attack triggered the mass de-registration of 46,812 accounts on the company's Bandai Channel streaming service without the knowledge or consent of those users, Tokyo's Metropolitan Police Department said.
The student admitted to the allegations, telling investigators he had written the initial source code himself before turning to ChatGPT to refine and complete it.
"I created the source code for the withdrawal process myself. Since the processing was taking a long time, I asked ChatGPT and completed it in a different programming language," he told investigators, according to the Asahi Shimbun.
He said he held no grudge against the company, explaining that he had targeted it simply because there were many accounts he could log into.
Even after Bandai Namco Filmworks took countermeasures and blocked his access, the student allegedly changed his IP address 30 times to continue sending the false commands.
The company was forced to suspend services for more than a month and issue refunds to affected members.
The following month, Bandai Namco Filmworks announced that personal information from up to 1.36 million accounts, including email addresses, account balances and payment methods, had potentially been leaked.
The company said no secondary damage, such as data being published online, had been confirmed.
"We take this situation very seriously and will continue to conduct regular checks and strive to prevent any recurrence," it said, according to Japanese media.
The student had taught himself to code as an elementary school pupil and told investigators he enjoyed analysing network communications.
He had already been arrested in June on suspicion of logging into the streaming service using another member's account, which led investigators to uncover the wider attack.
A senior Metropolitan Police Department investigator issued a warning following the arrest.
"Cyberspace is highly anonymous, and people may be tempted to commit crimes casually, but these actions can lead to grave consequences," the investigator said.
The case is among the first in Japan to involve the use of generative AI as a tool in a cyberattack, raising fresh questions about how easily such technology can lower the barrier for criminal activity online.
Read MoreJapanese woman arrested for sewing roommate’s mouth shut
Climbers rescued by helicopter from Japan mountain after bear encounter