A Predictive State? The Sri Lankan Context of Predictive Policing

195

The Sri Lankan state sector has shown gradual warming up to digitalization. Taking a lead from President Gotabaya Rajapaksa’s Digital Sri Lanka initiatives, there are ambitious plans to give the Sri Lankan state sector a much-needed digital makeover. However, there is one area of the state that has been sluggish to implement digital transformations, it is perhaps the most vital state institution, as it deals with the critical function of maintenance of law and order. If you haven’t guessed already, it’s the Sri Lanka Police. 

The Digitalization of Sri Lanka Police (SLP) has been a slow journey. SLP has 496 police stations. Of these stations, 492 are on a VPN network (4 Police stations are unable to be connected to VPN due to technical issues with Broadband connectivity). Digital use cases by the police are limited in application. 

They only use their email system to share Daily Police Messages (Bulletins, Notices by IGP or Senior officers etc.) but due to the COVID situation, the email system has now been expended to be used for divisional communications. There is an Admin/HR system that contains details of all 88,000 police officers. This includes their personnel records, leave information etc. Finally, individual police stations’ criminal data & traffic offence data is manually entered via web forms and updated to Police CRD (Criminal Records Division) for trend analysis.  

Despite being slow in digital adaption, SLP has recognized that Data Analytics is a vital part of modern policing. To assist in that the Criminal Records Division has set up local infrastructure, as well as a web server hosted in a managed data centre to host raw criminal data stats for analytics. Right now, it may be in its infancy, but many law enforcement agencies globally are pivoting towards Predictive policing.           

Predictive policing? 

Predictive Policing
(Image credits: Reveal)

Predictive policing uses computer systems to analyze large data sets, including historical and present crime data. In doing so, help decide where to deploy police or to identify individuals who are purportedly more likely to commit or be a victim of a crime. 

Place-based predictive policing, the most widely practised method. It typically uses preexisting crime data to identify places and times that have a high risk of crime (E.g., Hypothetically, Kotahena at 6:00 pm)

Person-based predictive policing, on the other hand, attempts to identify persons or groups who are likely to commit a crime — or to be a victim of one — by analyzing for risk factors such as past arrests or victimization patterns.

Good Idea or Bad Idea? 

Advocates argue that algorithms can predict future crimes more accurately and objectively than human police officers relying on their instincts alone. Further, predictive policing can provide significant cost savings for police departments by improving the efficiency of resource deployment to their crime-reduction efforts.

Opponents warn about a lack of transparency from law enforcement agencies that administer predictive policing programs. They also point to several civil rights and civil liberties concerns. This includes the instances that algorithms could reinforce racial biases in the criminal justice system. These concerns, combined with independent audits, have led some police departments in the United States, including in Los Angeles and Chicago, to phase out or significantly reduce the use of their predictive policing programs.

Notable examples of predictive policing projects

One of the earliest adopters of predictive policing was the Los Angeles Police Department (LAPD). They began working with federal agencies in 2008 to study different prediction algorithm approaches. Since then, the LAPD has implemented a variety of predictive policing programs. Notable examples include LASER, which identifies areas where gun violence is thought likely to occur, and PredPol, which calculates “hot spots” with a high likelihood of property-related crimes: 

“..about past offenders over a two-year period, using technology developed by the shadowy data analysis firm Palantir, and scores individuals based on their rap sheets. If you’ve ever been in a gang, that’s five points. If you’re on parole or probation? Another five. Every time you’re stopped by police, every time they come knocking on your door, that could land you more points. The higher the points, the more likely you are to end up on something called the Chronic Offender Bulletin, a list of people the data says are most at risk of reoffending and ought to be kept on close watch…” 

The New York Police Department (NYPD) also created predictive algorithms for several crime categories, including shootings, burglaries, assaults, breaking and entering, motor vehicles thefts, and armed robberies. Those algorithms are used to help assign officers to monitor specific areas. 

The predictive case for Sri Lanka 

Predictive Policing
Following the establishment of the military drone regiment, it has been working closely with the Police to enforce lockdowns in isolated areas (Image credits: army.lk)

One could argue that this tech seems a bit far off as a use case in Sri Lanka. Nonetheless, the building blocks are in place. The CRD already has a database in digital format hosted in the data centre, as well as the Police CCTV division, which has been in operation since 2010 (Incidentally this unit was created in 2010 as the brainchild of current President Gotabaya Rajapaksa).  

In August 2019 it was decided to upgrade the police CCTV system with advanced optics which would allow Police to identify the vehicle’s number plate clearly and get a clear view of the driver’s face. (i.e., enabling facial recognition). The Police with the assistance of the military’s drone regiment has already started using advanced drones. These are equipped with Zenmuse H20 optics with thermal and night vision capabilities to enforce COVID-19 lockdown in isolated areas.

The missing piece of this Predictive policing puzzle is the Facial Recognition Tech which can’t be deemed as inaccessible, as evident by Russian Apps like FindFace: A NtechLab App Launched in the mid-2010s, where the app allowed users to take a picture of someone and match their face to their social media profiles on Russian site Vkontakte (VK). Although NtechLab since shut down the consumer app,  the company pivoted its tech to the lucrative surveillance market. In January 2020 the company disclosed that it’s being paid at least $3.2 million for deploying its tools across the Russian capital. NtechLab CEO Alex Minin claimed, in an interview with Forbes, that it’s the biggest “live” facial recognition project in the world.  

Comedian John Oliver in an episode of ‘Last Week Tonight’ on Facial Recognition painted a target on another tech company ClearView AI. The company was offering Law enforcement agencies Face Recognition tech by mining online user data to build a vast facial recognition database. Although under a slew of lawsuits, the company has shifted to COVID- 19 related contract tracing efforts using its platforms in order to gain legitimacy.

Finally

Thus, the implementation plan is clear. Scale up the Criminal Records Division, which is the data analytics unit of the Sri Lanka Police. Upgrade the CCTV division with latest Facial Recognition Tech, tap into the presumptive eNIC Project for location/demographic related data. Then the recipe is ready for a Sri Lankan Predictive policing program

So, what does this mean to regular Citizen Perera?  Is it Orwellian? Perhaps, would it be Effective? Overwhelmingly, is there a chance of abuse? Yes, without the necessary legal provisions, regulatory oversight and privacy councils it would be. But if used well and true to its purpose, it would transform the Sri Lankan Law enforcement to an agile, digital-first, effective, efficient entity. But then, for whom will its bells toll?   

LEAVE A REPLY

Please enter your comment!
Please enter your name here