AI developed in Japan to recognise online posts by minors, for their protection

A joint research team from the University of Tokyo and CyberAgent Inc., an information technology company, has developed artificial intelligence technology able to recognise online posts by minors.

The system could help prevent minors from getting caught up in online crimes or other trouble. It was presented at a conference of the Japanese Society for Artificial Intelligence held in late May in Nagoya.

The team used 10 million posts on a social media site run by a CyberAgent subsidiary to train the AI to recognise content and commenters' ages.

As a result, the AI learned to recognise posts by minors at a 90 per cent rate of probability.

Minors tend to post about their school lives or club activities, using words like "homework."

The AI categorizes commenters by age based on these terms.

"We want to play a role in protecting minors on social media," said Fujio Toriumi, an associate professor of the Department of Systems Innovation at the University of Tokyo, who is an expert on computational social science.

The AI's analyses are premised on comments that are made public, so it cannot be used for private posts on apps such as LINE, a free communication app.

"It is important to make sure such communications remain secret, but we are in an era where we need to reexamine how to handle information that will benefit society," said Masatomo Suzuki, a professor of information law at Niigata University.