Are programmes that analyse security data in Los Angeles effective?

One of the current trends in terms of security is the wish to predict crime by using computer programmes that draw on data, mainly lawsuits. There are more and more police forces that are doing tests with such systems or are implementing them, but rather than the actual launching of these programmes, the most interesting aspect is assessing results and the analysis obtained by such an assessment. In this blog we have previously emphasised two experiences in France and Germany and, on this occasion, we add another from the USA.

Los Angeles Police Commission

The Los Angeles Police Department published, in March 2019, a report with a review of the application of three programmes of data-guided police action. In summer 2018, after a meeting during which some associations had revealed their concerns that these systems were generating among the general public the Board of Police Commissioners[1] of the city of Los Angeles commissioned the Office of the Inspector General [2] to study the matter and the report is the outcome of this.

The first of the programmes analysed is the so-called LASER (acronym of Los Angeles Strategic Extraction and Restoration –, and it involved two complementary actions: a system to assess and monitor chronic offenders based on scores that were reviewed periodically, and a strategy to detect hot spots where patrolling had to be prioritised.

The review of the LASER project has been quite critical, both regarding the chronic offender programme, and the locating of hot spots. It reveals problems due to the nature of assessment, discrepancies between initial objectives and those that follow with implementation, a lack of training material, differences in how the different areas execute the programme’s actions and difficulties to confirm effectiveness. And between the deficiencies of each of the two actions the most noteworthy are the lack of monitoring of people involved in the system of chronic offenders or the localising of hot spots near the very infrastructures of the police department and the time the patrols spent on these premises, being computed as patrols in these areas.

The second system analysed is Predpol, one of the most popular crime prediction programmes. It is based on historic data regarding crime and applies an algorithm to calculate the areas where a crime is most likely to be committed at a moment in time. Los Angeles police department only uses it to predict places where vehicles could be stolen or broken into, and there are 12-hour time bands. The review carried out by the inspector general’s office, was based on the patrolling implications rather than the functioning of the system of prediction and stresses that, on the one hand, results were contradictory in terms of the time dedicated to patrolling and the areas identified by the system and, on the other, that in most cases the areas of risk were only visited for less than a minute (meaning that the deterring effect was limited) and that the longest-lasting patrols were on police premises.

The latest programme used was ELUCD, a system that sent brief questionnaires to electronic devices (telephones, tablets or computers) based on where they were in the city of Los Angeles. Apart from some demographic questions (age, gender, race or address) it was asked if the person felt safe in his neighbourhood, if he trusted the police and if he / she felt secure with the police department. When the survey was commissioned a contract was not being considered, and this did not happen and the company did not provide the Police Department with the data gathered, and only sent some occasional weekly reports.

The report has had a certain impact on the US media (look at the footnotes) and, apart from detecting the difficulty to assess the effectiveness of the systems, has criticised the fact the effects of these programmes and systems on the general public have not been borne in mind, as this was one of the main criticisms that the study had brought about.

[1] This is the organ that that manages the Police Department, and is made up of five civilians and establishes polices and sets objectives. The head of the police manages daily operations and implements these policies and objectives.

[2]It is a service that supervises the functions of the city’s police department.

You can consult the report at:

http://www.lapdpolicecom.lacity.org/031219/BPC_19-0072.pdf

Two noteworthy news items emphasised by the report:

CNN – LAPD audit reveals dangers of high-tech policing

https://edition.cnn.com/2019/03/22/tech/lapd-audit-tech/index.html

Muckrock – Eight years in, LAPD can’t measure PredPol’s effect on crime

https://www.muckrock.com/news/archives/2019/mar/12/algorithms-lapd-predpol/

_____

Aquest apunt en català / Esta entrada en español / Post en français

Comentari / Comentario / Comment / Commentaire

Fill in your details below or click an icon to log in:

WordPress.com Logo

Esteu comentant fent servir el compte WordPress.com. Log Out /  Canvia )

Google photo

Esteu comentant fent servir el compte Google. Log Out /  Canvia )

Twitter picture

Esteu comentant fent servir el compte Twitter. Log Out /  Canvia )

Facebook photo

Esteu comentant fent servir el compte Facebook. Log Out /  Canvia )

S'està connectant a %s