(C) asahi.com

TOKYO, Japan - Japanese police are leveraging Artificial Intelligence (AI) to significantly enhance their crackdown on illegal part-time job solicitations, often referred to as "yami arubaito" (dark part-time jobs), which have been linked to recruiting perpetrators for serious crimes. The deployment of the new AI system has demonstrated a remarkable sixfold increase in efficiency compared to previous manual inspection methods.

AI System Targets X (formerly Twitter) Solicitations

According to the Yomiuri Shimbun on December 8, the National Police Agency (NPA) introduced and began operating an AI system in late July to automatically detect posts soliciting illegal part-time work, initially focusing on the social media platform X (formerly Twitter).

The results of the four-month trial, spanning from August to November, have been impressive:

The system identified 18,500 posts recruiting for illegal jobs on X.
The NPA subsequently sent warning messages to the authors of these posts.
This figure is six times the volume of posts that police officers could manually identify and address within the same timeframe.
The NPA is now planning to expand the use of this AI technology to monitor other social media platforms in a concerted effort to curb the spread of these illicit recruitment tactics.

 
From Manual Input to Automated Threat Assessment

The issue of illegal job solicitations via social media came to the forefront following a series of robbery cases that occurred repeatedly in the Tokyo metropolitan area since August last year. Investigations revealed that social networking services (SNS) were the primary tools used to recruit the individuals who carried out these crimes.

Perpetrators arrested in connection with aggravated robbery and trespass used multiple X accounts, employing coded language or slang terms such as:

"White matter" (화이트 안건): A seemingly innocuous term used to disguise criminal activities.
"Same-day, same-cash" (즉일즉금 or 当日即金): Highlighting the promise of immediate, untraceable payment to attract financially desperate individuals.
Previously, police investigators relied on a manual process, where they would search for illegal recruitment posts by entering crime-related keywords. Once identified, investigators would send messages warning the authors that "recruiting perpetrators constitutes an unlawful act."

However, the surge in high-profile robbery cases linked to these dark part-time jobs prompted the NPA to recognize the limitations of manual searching and approve the strategic adoption of AI.

 
Sixfold Jump in Enforcement Capacity

After a preparation period, the NPA implemented the AI-driven system in late July. This system involves an AI that has been trained to learn the specific slang and expressions used in illegal part-time job solicitations. The AI then classifies X posts based on their risk level, prioritizing those most likely to be criminal recruitment efforts. Investigators verify the high-risk posts before the warning messages are officially dispatched.

The NPA reported that the introduction of the AI detection system has boosted the average number of daily warning messages sent from 25 cases (manual average) to 150 cases (AI-assisted), a sixfold increase. The 18,500 cases identified over the four months from August to November already far surpass the total of 9,000 cases recorded for the entire previous year.

The police emphasize that the warnings are proving effective, as many illegal job postings are being deleted shortly after the messages are sent. Given this success, the NPA plans to extend the AI monitoring to platforms like Facebook and Threads to comprehensively block illegal recruitment activities across the digital landscape.

The integration of AI into social media surveillance marks a significant technological step for Japanese law enforcement, shifting from reactive investigation to a more proactive and high-volume deterrent against crimes orchestrated through anonymous online recruitment.

[Copyright (c) Global Economic Times. All Rights Reserved.]