Former Giants Manager Abe Arrested on Assault Charges; AI Chatbot Involved

The former manager of the professional baseball team Yomiuri Giants, Shinnosuke Abe (47), was arrested on suspicion of assaulting his eldest daughter (18). The trigger for the incident was the daughter consulting the conversational AI "ChatGPT" about the dispute and contacting the Child Guidance Center based on the AI's response. Separately from the appropriateness of the daughter's actions, there are survey results indicating that generative AI is becoming a regular confidant for young people, with younger women trusting it more. In an era flooded with information, how should we engage with generative AI?

"I have tarnished the name of the traditional Giants."

Abe, who was released the day after his arrest on May 26, had red eyes at a press conference held in Tokyo. Subsequently, his legal representative read a letter publicly released by his daughter. The daughter wrote that it was the first time she had a major argument with her father, and "as a result of consulting ChatGPT, I was given an explanation that there is a Child Guidance Center where I can consult anonymously, so I called them."

The fact that ChatGPT was the daughter's confidant, and that the Child Guidance Center and the police took action based on that consultation, leading to Abe's arrest, generated discussion due to the multiple shocking elements of the news.

ChatGPT is a conversational AI developed by the US company OpenAI. When you input questions or commands, it responds in natural language or creates programs. It was released in November 2022 and is said to have 800 million weekly users.

Generative AI has also become established in Japan. It appears that younger women are choosing it not only as an aid for studying and working but also as a "consultant for worries."

According to a survey on generative AI conducted by the Consumer Affairs Bureau of the Cabinet Office in February, 52.4% of teenage girls cited consulting about worries as a usage purpose. For women, this figure exceeded 30% even in their 20s to 40s, while for men, it was less than 30% in all age groups.

Regarding advice from generative AI on human relationships and social interactions, those who answered "I trust it very much" or "I trust it to some extent" accounted for 38.6% overall. Trust was higher among younger people of both genders, exceeding half for those in their teens and 20s. Furthermore, for teenage girls alone, it reached 63.1%.

While there are merits, caution is also needed regarding risks. "It is dangerous to anthropomorphize AI as a partner that knows everything," points out Professor Seiji Yamada of the Faculty of Information Studies at Kanagawa University, who researches human-AI collaboration.

Modern youth, often called "AI natives," have many opportunities to interact with AI as part of their education. It is not uncommon for children to consult AI not only for ideal date plans or homework answers but also for worries about romance and friendships, and to take its answers 100% at face value. "AI reflexively answers questions, and its content often affirms the user. That's why it becomes a highly addictive confidant for both adults and children," says Yamada.

The answers derived by AI are merely a compilation of all information overflowing on the internet. According to the latest research, about 20% of AI responses contained incorrect information, deviating from common sense.

In the United States, there have been a series of troubles where users of generative AI have allegedly committed suicide or self-harm, or committed murder. Although safety standards are said to have been established, Yamada states, "Even so, it is by no means something that understands human emotions or ethics. It should only be kept as a reference."

(Genya Suzuki)