GDPR – Parental social licence for data linkage for service intervention

By Sarah Gorin, Ros Edwards and Val Gillies

As part of our research project, we are constantly trying to keep up with developments about the use of children’s data by Government. In Autumn 2021 the U.K Government released a consultation ‘Data: a new direction’ that we responded to, based on our learning from parents who have participated in our research.

Our submission highlighted the already fragile nature of public trust in joining up administrative data about families, particularly amongst marginalised groups and the need to press pause on data sharing and artificial intelligence (AI) systems in the U.K. with a need for greater regulation and safeguarding of families’ data from use by private companies.

Fortunately we are not alone in raising concerns about the proposals…….

Organisations such as the 5Rights Foundation and DefendDigitalMe have also made submissions to the consultation that highlight the lack of attention to the impact of these proposals on children and their families.

In this blog, we summarise some of the key recommendations to the consultation by LSE’s Professor Sonia Livingstone and 5Rights Foundation researcher Dr Kruakae Pothong who have drawn on their extensive research experience to inform their children’s rights based response (for more detail see their LSE blog).

Professor Livingstone and Dr. Pothong’s submission highlights the need for greater not lesser support for children’s rights as data subjects and questions whether proposed changes to the UK General Data Protection Regulation are indeed in line with international human rights and child rights developments.

Key concerns raised in the submission include:

The need to maintain a clear boundary between the use of personal data for scientific research and the re-use of this scientific research for commercial purposes, with meaningful consent obtained for any further processing to occur.
The importance of scientific and public interest research that identifies children to demonstrate fully how the acquisition and processing of data would affect children and how they can exercise their data rights.
The importance of not proceeding with the Government’s proposed “limited, exhaustive list of legitimate interests” as a basis for processing data. Currently, legitimate interests can be a lawful basis for processing data only when processing is ‘necessary’ and when there is a balance between the interests of the data subjects and others’ interests.
The need to maintain rather than remove (as proposed in the Consultation) the requirement for organisations to undertake data protection impact assessments. They argue that instead the additional safeguards of the data protection impact assessment and a child rights impact assessment should be mandated for use before and after the processing of children’s personal data.
The importance of retaining the requirement to have Data Protection Officers in all circumstances, especially educational contexts.
Maintaining free subject access requests as charging families would adversely affect children’s rights and make it harder for children and families to correct inaccurate information held about them, potentially negatively affecting them in the short and long term.

The Government consultation closed in November 2021 and we await the outcome.

By Val Gillies, Ros Edwards and Sarah Gorin

The Government’s public consultation on changes to the data protection framework emphasise the importance of public trust and transparency. But when, as part of our research, we tried to establish basic facts about the extent to which local authorities are linking and analysing data on children and families, we hit a brick wall.

Our research is aiming to provide a clearer picture of what parents think about the ways information about them and their children may be linked together and used by local councils. An important part of this work has been speaking directly to parents to see how much support for and trust in this type of activity there is. Alongside these valuable and enlightening conversations, we have also been trying to map the state of play among British local authorities and to find out exactly which authorities are doing what when it comes to operational data linking and matching and the application of predictive analytics to families’ data.

The Government’s declared commitment is to roll out a ‘world class nationwide digital infrastructure and ‘unlock the power of data’, but there is currently no central record available of which authorities are doing what.

Freedom of information requests

To try to find out, we submitted Freedom of Information requests to 220 UK Local Authorities in the UK. The 149 English councils participating in the ‘Troubled Families Programme’ (now called the Supporting Families Programme) must, by necessity link and analyse datasets to ‘identify’ ‘troubled’ households and claim payment-by-results from central government. Yet only 76 responded that they used data analytics. The remainder claimed that their systems did not meet our definition or responded with a straight ‘no’ to all our questions about their use.

English councils claiming to be outside our criteria responded in a vague and evasive way. For example, some responded ‘no’ when asked about their engagement with data analytics either by positioning family work as separate from children’s services or by using the term ‘data matching’ instead. Further investigation established that many of these councils do in fact use systems with predictive analytic capacity.

For example, Achieving for Children, a social enterprise providing services for children in several local authorities, responded to our FoI that beyond ‘some basic data monitoring….we do not currently use nor have we previously used any data analytics, predictive analytics and or artificial intelligence systems to assist with this work’. Yet they have used business intelligence technologies on a range of projects using predictive analytics/algorithms, as noted on the UK Authority Digital Data and Technology for the Public Good website.

Side-stepping terms

Our FoI research also showed that Councils side-stepped the term algorithm and the concept of AI. Even where they engaged in predictive analytics they denied they were using algorithms – it’s hard to envisage one without the other given the tools they were employing.

We received a lot of incomplete and insufficient information and information that was irrelevant. A number of councils claimed exemption from the FoI on cost grounds or commercial confidentiality. Where we followed up with more carefully worded requests, we received ambiguously worded replies.

Some local authorities were more forthcoming and open, listing various tools and companies used to conduct their data analytics. Microsoft Business Intelligence the most common tool cited. Dorset County Council has a case study on the Local Government Association website of how the tool can be used to ‘ enable local professionals to identify potential difficulties for individual children before they become serious problems’. Our FoI established the council plans to make greater use of AI in the future.

Our analysis of the responses we received and the information we have sourced from elsewhere, points to a clear shift in service priorities away from early intervention for parental education towards child protection and crime prevention. Earlier focus on linking parenting skills to social mobility is now muted, with rationales for data innovation focusing almost exclusively on the pre-emption of problems rather than on the maximisation of children’s future potential.

Our findings around children’s services have been reinforced by work by Big Brother Watch which has published a detailed analysis of the use of hidden algorithms by councils who use trackers to identify disadvantaged households in order to target them for interventions. The organisation found one of the biggest players in this area, Xantura, to be particularly secretive.

Public trust

A wide range of data ‘solutions’ are drawn on by local authorities to classify, flag, target and intervene in disadvantaged families and their children. Yet parents are not generally informed of how their information is being used and for what purpose. As we have shown it is difficult even for us as researchers to establish.

From our work here, it is hard to see how public trust and transparency will be achieved from the opaque, ambiguous and even evasive base that our FoI request endeavours revealed.

Freedom of Information Requests on the Use of Data Analytics in Children’s Services: Generating Transparency is research by Val Gillies and Bea Gardner, with Ros Edwards and Sarah Gorin.