Will algorithms predict your future?

19 November 2018

algorithms

A report from Cardiff University reveals the extent to which public service provision is now being influenced by data analytics.

The study, Data Scores as Governance, represents the culmination of a year-long research project compiled by the School of Journalism, Media and Culture’s Data Justice Lab.

Researchers studied data systems of government services across the UK to assess the extent to which decisions about individuals are being made by data scores and algorithms. They also conducted interviews with employees working in the field to find out how information gathered about people they come into contact with is used and distributed.

Their findings indicate that data collection and data sharing across local councils and government departments is now widespread.

Data Justice Lab Co-Director Dr Arne Hintz said: “We’ve become accustomed to our data being used online to decide which adverts or social media posts we should see. But for the first time, our research shows the extent to which our data is being used to decide how we are treated by public services.

Arne Hintz

Most people would know what a credit score is and where online they could check their score. But very few people know how data about them is used to produce scores in public services and there are few, if any, avenues to find out about it.

Dr Arne Hintz, Senior Lecturer (Teaching and Research)

Co-director Dr Joanna Redden added: “In most cases, we found data systems had been implemented in response to austerity measures as a way of prioritising resources.”

Assigning 'risk scores' around vulnerability and crime, in particular, is a growing trend, the report’s authors note.

For example, Avon and Somerset Police’s Qlik Sense system was first piloted in 2016 and now has over 30 applications across teams. It serves as both a performance assessment tool and a predictive policing tool. The platform creates profiles of any individual that has interactions with officers, before assessing their level of risk.

Local authorities also routinely collect information and use this data to prioritise the level of support an individual receives. Bristol City Council’s integrated analytical hub combines data of “all social issues” across the city for children and families to provide a “holistic understanding” of their needs. The system is now using predictive modelling to prepare for “future trends”. The Camden Resident Index is run by Camden Council, which was the first local authority to have implemented a master data management system, to allow for a “single view of a citizen”.

The full findings of the report will be presented at London’s Central Hall Westminster today.  (Monday 19 November 2018.) An interactive online map has also been developed to allow policymakers, journalists and members of the public to explore how different organisations in the public sector use data analytics in their decision-making.

Co-Director Dr Lina Dencik added: “There are no standards in place for how data systems are implemented across local authorities. Some are developed in-house whilst some are out-sourced with no set requirements for auditing consultation or impact assessment.

Lina Dencik

While the report will document to some degree how data collection systems and data scores work, it is still unclear ultimately who these systems serve and to what purpose. Data scores are becoming a way to categorise citizens, allocate services and predict future behaviour.

Dr Lina Dencik, Reader

The conclusions come as Professor Philip Alston, the United Nations Special Rapporteur for extreme poverty and human rights, finishes a two-week inquiry into rising levels of poverty and hardship across the UK, which includes a focus on algorithms in welfare.

His report will feature testimony gathered by the Data Justice Lab from professionals on the frontline and those working on poverty, welfare and citizen rights, who share their concerns about the digitisation and datafication of public services.

For more information visit the School webpages.