Greater London Authority
Preventing adverse impacts of data sharing for people ‘at risk’ of violent crime
The challenge
Since 2015, violent crime in London had increased in almost every year. To curb this trend and protect Londoners, the GLA’s Violence Reduction Unit adopted a public health approach and focussed on the social determinants of violent crime.
This approach relies on the sharing of personal data to prevent crime. But sharing this type of data could have disproportionate impacts - both positive and negative - on already marginalised communities. It is therefore essential to understand what people ‘at risk’ of being involved in crime, either as victims or perpetrators, perceived to be at stake and to determine their willingness to support the sharing of personal data.
Our approach
We conducted in-depth interviews with young people from diverse ethnic backgrounds, aged 16-30 years-old, living in selected wards and with profiles known to be ‘at risk’ of being involved in crime. Our interviewees included people:
with a background of truancy, temporary or permanent exclusion from school
aged under 18 and N.E.E.T.
involved illegal drug use
who had been in local authority care
with a Conduct Disorder or Oppositional Defiant Disorder
with family or friend involved in violence
in friendship groups that commit crime and/or engage in anti-social behaviour
To elicit attitudes about such a complex topic, we created vignettes that depitcted sharing different types of data, for different purposes, betwen different agencies that could help reduce violent crime.
Insight
We found that ‘at risk’ communities understood ‘crime prevention’ as efforts to prevent young people from ever getting involved in crime - not in terms of protecting themselves from becoming victims of crime.
If data sharing could achieve this benevolent purpose, they supported it. But they also needed reassurances about whether there was a genuine need to know, how effective data sharing was, which organisations knew what about them, their right to access their data and to ‘disappear’, and redress in case of data mishandling.
People accepted the sharing of personal data on an case-by-case basis but opposed mass surveillance.
They were highly sensitive to the risk of being further criminalised through data sharing. They discussed pitfalls linked to racial profiling, prejudice and bias (including those built into algorithms), human error, data breaches and improper sharing, the retention of obsolete data, and the loss of privacy, among others. Some also expressed concerns about the inability of the most vulnerable (e.g. with cognitive impairment) to provide informed consent.
They broadly trusted schools and health services to use personal data safely but distrusted the police.
We also identified differences in attitudes based on age, ethnicity and socio-economic group.
Impact
Our findings and recommendations informed the work of the GLA’s Violence Reduction Unit. In particular, it is used by key statutory sector organisations (e.g. Metropolitan Police, schools, NHS, ICO, etc) involved in data sharing, providing guidance on the types of personal data being shared, the protocols for data sharing, and public communications about why and how personal data is shared to prevent violent crime in London.
