What is Traffic Jam?

Traffic Jam is a facial recognition technology (FRT) tool developed by Marinus Analytics to assist law enforcement to fight organized crime, stop human trafficking, and find missing persons. The software scrapes data from publicly available websites to build a database of images, phone numbers, and location data. Traffic Jam uses Amazon Web Services’ Rekognition software to match a suspected victim’s photo with missing persons advertisements, social media content, and to determine if the victim’s face appears in any current or deleted sex or human trafficking advertisements.

Initially developed in 2014, the software purports to have an 88% success rate in identifying victims. According to its website, Traffic Jam is currently being used as an investigative tool by law enforcement in the United States, Canada, and the United Kingdom.

Use by Police Agencies in Canada:

Traffic Jam’s FRT software has been used by the Ottawa Police Service (OPS) in the context of human trafficking investigations as discussed in the case of R v. Ahmed. In this case, the OPS used Traffic Jam searches to uncover a number of advertisements posted on Backpage, a classified advertising website, for escorting services featuring the two complaints, who were aged 14 and 15, and their abusers. The Traffic Jam searches were then submitted as evidence to convict the defendants of human trafficking.

In the court proceedings, the Defence challenged the admissibility of the Traffic Jam searches as evidence for being unreliable and argued that the Crown’s submissions lacked information concerning how the software worked, who designed it, and what type of advertisements it captured. Worryingly, the court noted that the lead detective’s knowledge of Traffic Jam and its functionality was limited. Additionally, during oral testimony, the Crown’s lead detective revealed that they had used the software in additional investigations with other Canadian police services, such as the York Police Service.

Ultimately, the court rejected the Defence’s claims and permitted the Traffic Jam evidence by cross-referencing the search results with additional evidence provided by the Crown, such as proof that the defendants had accessed Backpage, identical photographs contained in the advertisements that were found on the defendant’s phone, and time stamps of the advertisements’ publication.

Significance:

This case demonstrates that law enforcement agencies in Canada are employing a variety of facial recognition tools to assist with their investigations. While this software may have socially beneficial uses, it remains unregulated and open to potentially harmful uses by police agencies. For example, the use of FRT facilitates function creep, whereby information collected for one purpose – for example, finding victims of trafficking – leads to other less socially-defensible or more debatable uses. This can lead to particularly harmful consequences including the surveillance of equity-deserving populations, including sex workers. Moreover, without robust safety standards in place, such issues may be exacerbated by law enforcement agents’ unfamiliarity with FRT’s functions and capabilities. Finally, Traffic Jam is supported by Amazon’s Rekognition software which, researchers have demonstrated, performs worse at distinguishing female appearing faces than male and at distinguishing between darker faces than female faces.