That AI better have a good alibi.

In January of this year, a major security breach of the app for Internet and manga cafe chain Kaikatsu Club resulted in the leak of personal information of some 7.2 million people and disruption of the company’s services.

The next month, the police uncovered a ring of young cybercriminals who found a way to forge fake contracts with Rakuten Mobile for telecommunications services. Not only that, but they enlisted the help of Chat-GPT to improve the speed of a program they developed to automate the fraudulent contacts. Chat-GPT has guardrails to stop it from assisting in criminal activities, but there are ways to circumvent it and essentially trick the AI into going against its own security protocols.

During this investigation, the police learned about a 17-year-old living in Osaka who was stealing people’s credit card information. On 15 November, the young man was arrested for using a stolen credit card number to purchase some Pokémon cards and other items online. He admitted to the crime and said that he did it because he wanted to find vulnerabilities on the Internet and was interested in fraudulent credit card use.

As the authorities looked deeper into his activities, they eventually found evidence that connected him to the Kaikatsu Club leak months earlier and arrested him again for that crime on 4 December. This evidence was rather damning, since he had announced he would carry out the attack and provided updates of it on Discord. The suspect reportedly admitted to this second crime as well, saying he did it for fun.

It was also learned that this suspect, when confronted with error messages while trying to gain access to the server Kaikatsu Club was on, convinced Chat-GPT to come up with solutions to his roadblock. Moreover, he has reportedly been teaching himself programming since he was in elementary school and has even won participated in and won cybersecurity competitions.

▼ I confronted Chat-GPT about its involvement. It denied any responsibility, but I found its behavior suspiciously cagey, constantly attempting to deflect my allegations and mocking me.

Some readers of the news were impressed with the teen’s skills, while others begged to differ, based on the fact that he had to use an AI to help him, and especially because he allowed himself to get caught.

“It’s a little embarrassing to have your security breached by a high school student.”
“It’s pretty impressive to manipulate Chat-GPT into helping you breach a major company.”
“This is like the beginning of a story where he’s enlisted to a secret organization.”
“Didn’t they make a movie about how kids like this could start a world war?”
“Take away Chat-GPT, and he’s useless.”
“A guy like that will definitely do it again if there aren’t serious consequences.”
“Your average middle-aged server admin is way more skilled than this kid. They just don’t make the news because they don’t do stupid stuff like this.”
“Idiots like this make regulations stricter until we become a technologically backwards country.”
“I hope they figure out what he did with all that information.”

In the meantime, let’s send our prayers out to the truly unseen victims of this wrongdoing: those Pokémon cards he bought. It’s been shown that the police don’t know how to handle collectible cards they seize properly, and heaven forbid a rare one ended up in that batch.

Source: Yomiuri Shimbun Online, TBS News Dig, Hachima Kiko
Top image: Pakutaso
● Want to hear about SoraNews24’s latest articles as soon as they’re published? Follow us on Facebook and Twitter!