Police Use AI to Identify 18,500 ‘Dark Part-Time Job’ Recruiting Posts, System Significantly Expands Deterrence Efforts

The Yomiuri Shimbun

16:28 JST, December 8, 2025

The Metropolitan Police Department has used artificial intelligence to automatically detect posts used to recruit for “dark part-time jobs” found on X and sent warnings to about 18,500 such posts over the four months from August to November, it has been learned.

The introduction of the technology expediated the police’s task sixfold compared to work done manually by police officers. The MPD will utilize the system to scrutinize other social media platforms, including Meta’s short-form posting app Threads, to prevent young people from becoming involved in crime.

Such postings have been used to hire perpetrators in a spate of robbery incidents that have occurred in the Tokyo metropolitan area since August 2024. The masterminds, who were recently arrested by the MPD and three prefectural police forces on suspicion of robbery and other charges, allegedly used multiple X accounts to recruit perpetrators with phrases such as “legitimate opportunities” or “same-day payment in cash.”

The MPD had used keywords to manually search for dodgy posts and then sent a warning reading “Recruiting perpetrators is illegal” to suspected illegal postings.

Following a surge in illegal job-related robberies last year, the MPD introduced AI in late July in which the system learns slang and phrases used in illegal job postings, extracts X posts and classifies them by risk level. Warnings are sent after police check the contents.

The AI system increased the average number of warnings the police sent daily from about 25 to 150. In the four months since August alone, the number reached 18,500, surpassing the roughly 9,000 sent in all of last year, according to the MPD.

Posts are often deleted after receiving a warning, it said.