AI berubah memerangi penyeludupan manusia

photo Kredit: Ira Gelb/Flickr

AI is transforming the human rights space and encouraging civilian participation in the fight against human trafficking.


Combatting human trafficking is like a game of cat and mouse. For every progression in investigation techniques, traffickers devise equally novel methods to evade capture and continue profiting from their operations.

But artificial intelligence and machine learning are changing the game. New technologies are altering the way investigations into human trafficking are being conducted. These technologies are triggering a shift to citizen-driven investigations. Instead of investigators coming up against an information wall, modern human trafficking investigators have access to masses of information at the click of a button. The new challenge lies in interpreting that data.

A new app to identify victims of human trafficking

In February, researchers in the US revealed they had developed an AI solution that could assist in uncovering the location of sex trafficking victims. The AI engine, named Hotels-50K, relies on a database of more than one million images taken from 50,000 hotels around the world. When a photo of a victim is identified, the engine compares the image to those in its dataset to pick out the location of the specific hotel room.

The engine uses features like furniture, colour schemes, artwork on the walls and bedding to identify the hotel in question. The engine reportedly has a success rate of identifying the correct hotel 24% masa. Given that three out of four underage victims of sex trafficking are advertised online, featuring photos frequently taken in hotels, the technology could become an essential tool in the fight against sex trafficking.

When combined with other AI solutions, the technology becomes even more valuable

Hotels-50K is useful for identifying the location of a known victim from a photo. Other AI and machine learning solutions can identify potential victims. Stylometry, for example, is a branch of machine learning which identifies language patterns in advertisements to identify an author.

Once a trafficker is detected, stylometry can be applied to reveal other advertisements written by the same author. It analyses word repetition, punctuation ad emoji use to identify the same human traffickers advertising multiple victims, even when they advertise the victim under a different phone number and email address.

During testing, the algorithm correctly identified an author from a data set dalam 90 daripada 91 kes-kes.

Traffic Jam, another AI engine, helps identify runaways and missing people advertised online through facial recognition software and geospatial analysis. IBM, in conjunction with Western Union and European authorities, maju a cloud-based information sharing system to detect suspicious payments that could have been made for human trafficking victims. This information is being employed to spot patterns and trends which, the computer giant hopes, will prove useful in predicting human trafficking payments before they occur.

These tools could be utilised to detect children for sale on social media in Indonesia, sex workers trafficked through Singapore dan trafficked Vietnamese brides di China.

ASEAN is also employing technology to combat human trafficking

Malaysian non-profits Tenaganita and Change Your World launched an app last year which allows users to anonymously report incidents of human trafficking and send photos to investigators. The app, Be My Protector, is the region’s first app dedicated to combating human trafficking. Since last April, it has received 300 reports from both victims of human trafficking themselves and from the general public offering information on people they suspect of falling victim to human trafficking.

Tenaganita’s director, Aegile Fernandez, bercakap dengan ASEAN Today. “All our cases go into the database, you know, so it is easier to look at the violations, to look at… the perpetrators," dia berkata, menambah, “we know the hotspots.”

At the moment Tenaganita has to use humans to coordinate with the victims to determine their location based on noticeable landmarks that appear in the photos. Walau bagaimanapun, Aegile Fernandez acknowledged that she hoped the organisation would be able to expand the app’s capabilities in the future. “In our conversations… we are saying now can we go further and use it for the [whole] ASEAN region?"

Aegile Fernandez, Director of Tenaganita
photo Kredit: YouTube / Screengrab

The technology is part of a wider shift in policing human trafficking

The rise of AI solutions is transforming the way investigators approach human trafficking investigations. Apps like Be My Protector and AI engines like Hotels-50K are encouraging civilian participation in human trafficking investigations. These technologies are harnessing the power of civilians and their smartphones to collect data on behalf of the authorities.

As a result, the role of investigators is also evolving. The newest challenge for many investigators is not necessarily finding the data but understanding it. Once again, AI is offering a solution. Technologies which allow investigators to quickly filter through the masses of citizen-gathered data are streamlining investigations.

Funding remains the biggest hurdle for ASEAN. “We are trying to seek funding in order to come up with [a way of] taking it further,” Aegile Fernandez said. To meet this funding dearth and unlock the potential of AI in the fight against human trafficking, business engagement is essential. The technical expertise and finance the private sector can bring to the fight will be invaluable.

Like in finance, insurance and agriculture, AI is set to disrupt the human rights space in ASEAN. The technology is already available. The sector is just waiting for the opportunity to unlock its potential.