Youth justice: academics argue against algorithms and actuarial assessments

By Stephen Easton

Friday October 27, 2017

Two academics are demanding the New South Wales police open up about their efforts to prevent future crime using an opaque predictive algorithm, which they say is “inadvertently encouraging unlawful police conduct” and leading to counter-productive harassment of young people, especially those from Aboriginal backgrounds.

A report published by the Youth Justice Coalition NSW this week is highly critical of suspect target management plans. The researchers found they are “disproportionately” applied to Indigenous people under the age of 18 and they say this is leading to outcomes that are contrary to the wider objectives of the youth justice system.

“Once a person is on the STMP they are repeatedly targeted by police, including being stopped and searched and visited at home regardless of whether they have committed a violence offence or minor offences like shoplifting or graffiti or even if they are just known to police but have never been convicted of an offence,” according to the researchers.

The report was written by principal investigator Dr Vicki Sentas, a senior lecturer in law at the University of New South Wales who jointly runs the Redfern Legal Centre Police Powers Clinic, and Camilla Pandolfini of the Public Interest Advocacy Centre.

“This type of heavy-handed proactive policing is very damaging to the relationship between young people and the police and we believe it undermines key objectives of the NSW Justice system, including diversion, rehabilitation and therapeutic justice,” Sentas said in a UNSW statement.

An algorithm helps police calculate a person’s future risk of offending and classify them as either extreme, high, medium or low risk. This, the report explains, is coupled with a “policing program that guides police interaction with individuals who are subject to the program” but the police have kept a tight lid on information about how this works.

There are often “discriminatory assumptions” behind the interventions with targeted youths and “serious issues of procedural and substantive justice” have emerged, in Sentas’ view.

“It is particularly concerning that the policy appears to disproportionately target young Aboriginal people,” she added.

The new approach deliberately increases how much contact young people have with the cops but the researchers were not able to obtain any evidence that this has led to a reduction in crime.

“If NSW Police say the STMP prevents crime, then show us how – make the data publicly available and let the Bureau of Crime Statistics and Research (BOCSAR) assess it,” said Sentas.

She argues there is unjustified secrecy around the program and that it should be opened up to public scrutiny, and proper independent evaluation.

The main recommendation of the report is that the police should stop using the opaque algorithm-based program to target children. Those believed likely to offend should be put through “evidence-based prevention programs that address the causes of reoffending or locally based programs developed in accordance with Just Reinvest NSW” instead, it argues.

As for police acting unlawfully: “In some instances, the exercise of police search powers in relation to a young person on the STMP have been found unlawful by the courts. The STMP may be inadvertently diminishing police understanding of the lawful use of powers … and thereby exposing police to reduced efficacy and civil action.”

STMPs are also used to target repeat domestic violence perpetrators, which involves “conducting overt checks and covert surveillance to ensure they are abiding by the law and not breaching their domestic violence orders” according to a 2016 government statement.

“Offenders do not know when police may be watching them or may next approach them, making them a powerful deterrent to violent behaviour,” it adds.

The thrust of this week’s report is that young people are being targeted in a similar way but for far less serious crimes like graffiti, and this is in fact breeding resentment and perhaps increasing their likelihood of continuing with a life of crime.

One of several case studies included concerns a 16-year-old referred to as “Dean” who has been caught for graffiti and trespassing in the past. He was stopped and searched 10 times in three months and a further 13 times over the next seven months.

“Dean’s police records indicate a range of justifications recorded by police as grounds for stops and searches,” according to the report.

“This included that persons of interest who wear Nautica clothing are ‘known to commit criminal damage offences’ (graffiti); that young people who get on the last carriage of a train and wear Nautica are known to commit criminal damage; and that Dean was with a group of young people.”

A criminal law firm raised concerns about the program in 2015, describing its use of “actuarial risk assessment tools” as a “major flaw” due to serious uncertainty over their accuracy in criminology.

According to the blog post by LY Lawyers, risk factors taken into account include whether the person is Aboriginal or Torres Strait Islander, their gender, school attendance, and whether their “peers” or family members have committed offences, along with their own criminal history and information about alcohol or drug use.

“It is inconceivable that a court would consider facts such as those above as a basis to increase someone’s sentencing,” the firm states. “For example, a judge would not increase an offender’s sentence just because he was male, had a previous record of skipping school, and has a mother who is an alcoholic. However, the police use these risk factors to legitimise their lists and the subsequent deprivation of personal liberty.”

“Another issue with actuarial risk assessment instruments is that police monitoring is based on probabilities and not on any actual crimes being committed. The judiciary process does not allow for someone to be accused of a crime based on probabilities. For example, a search warrant would not be approved without evidence or probable cause. However, the Suspect Target Management Plan allows for individuals to be targeted without a crime having been committed as a precondition to the removal of liberty.”

The researchers called for the program to be investigated by the Law Enforcement Conduct Commission. Chief commissioner Michael Adams reportedly said it could be problematic at face value and is discussing whether an inquiry is warranted with his fellow commissioners.

The report argues it should be mandatory for police to tell people (of any age) when they are being targeted due to suspicion of future crime, and that anyone classified as low risk should not be subject to STMP at all. The authors note that the STMP has proceeded so far with very little public accountability, and call for more transparency.

“Findings in this report are based on i) available quantitative data on program participants, ii) de-identified case studies drawn from interviews with lawyers, iii) publicly available guidance given to police on STMP operational procedures, and iv) analysis of case law and legislation,” they explain.

“The research has been limited by the lack of publicly available information on the STMP and the absence, to date, of scrutiny and oversight of the program. By adopting a mixed methods approach, the report is able to make robust preliminary findings and identify areas for further investigation.”

Comparisons with the classic sci-fi novella The Minority Report or its film adaption can be predicted with reasonable certainty. But just as the book’s protagonist finds out, the future is never entirely certain.

About the author
Inline Feedbacks
View all comments

The essential resource for effective
public sector professionals