Data dangers
The ongoing drive from successive governments to store and distribute ever increasing amounts of information about children and their families ignores the significant harms which can be caused.
How we are viewed and treated by professionals and services differs depending on what story the databases they access tell them about us, our ‘digital selves’. There are many problems arising from this. The lack of context or mitigating factors, the focus on deficit rather than strength, the way people are kept trapped in past difficulties, and how unreliable the conclusions are all reason to object strongly to this approach. Combined with a punitive system this leads to children and families who have suffered hardship being treated in negative ways, which multiplies the hardship and harm done to them.
What is even more unjust is that children and families are judged and treated not as individuals but on the basis of judgements made about ‘people like them’. Even worse is that these judgements are often made on very poor grounds, and misinterpretation of facts and statistics.
The harm and injustice caused has been significantly escalated by the linking up of databases across government – in effect creating a super database – and the use of automated decision making across the public sector. Any hardship or difficulty in one area of life – including injustices not of individuals making – then fans out across all.
The relationship between individuals and the state has become one where the state has all the power, holds all the information, makes all the decisions. Individuals are unable to access or to challenge incorrect information.
“ …in too many cases, the public are neither served nor protected by the increasingly complex and intrusive holdings of personal information invading every aspect of our lives.”
Professors Ross Anderson and Ian Brown, Terri Dowty, Drs Philip Inglesant, William Heath, Angela Sasse
Database State (2009)
“We know people are already being harmed at individual, community and societal levels by data systems. Data is being used in ways which exploits, discriminates, invades privacy, surveils, injures, manipulates and excludes.”
Professor Joanna Redden
Data Justice Lab: Video about our book, Data Justice (2024)
“Datafication becomes an integral part of the unequal power dynamic between parents and professionals, further shifting the balance towards professionals, damaging fragile trust and engagement.”
Professors Sarah Gorin, Rosalind Edwards, Val Gillies and Dr Helene Vannier Ducasse
‘Seen’ through Records: Parents’ Access to Children’s Social Care Records in an Age of Increasing Datafication (2024)
Linked databases
With public and political discussion focused on the need to ‘share’ information the extent to which this already happens is obscured. There is little understanding of the impacts and power of linked databases
With data and flags being shared across systems this can lead to negative treatment from other services. Linking up of systems means that difficulty in one area of life can impact on another. Being known to children’s services can increase the chance of intervention and assessment, itself a difficult and intrusive experience for both child and family, and has been shown to make teachers expect less of children and police to make assumptions about them.
The sharing of information and the loss of privacy can become barriers to both children and families asking for help, such as when families fear being pulled into child protection assessments or punitive policing.
“It is all families who are implicated in across-the-board data sharing, data linking, and application of predictive modeling, not just the families with “vulnerable children.””
Professors Rosalind Edwards, Val Gillies and Dr Sarah Gorin
Cambridge University Press Data linkage for early intervention in the UK: Parental social license and social divisions (2021)
Is it fair to keep records of everything?
How can children have a fresh start if data follows them?
Data can misrepresent
- Context is stripped out.
The context to behaviour is important. For example, a child who loses a parent may well experience difficulties going to school. Such context is lost, especially once ‘flags’ are shared.
- Flattening effect.
For example, a parent who smoked cannabis after a child were asleep might be recorded the same as a parent with a heroin addiction, even though these two present very different safeguarding challenges.
- We have no ability to challenge the inferences made about us which are drawn from data.
For example, the Home Office has claimed that there are links between low school attendance and involvement in serious youth violence. This could change the way that a child is treated by police, or other public bodies. The fact that the link is not statistically robust does not unfortunately prevent them from drawing these negative conclusions about a child. There are similar risks with links being drawn between neurodivergence and involvement in the criminal justice system. Professor Sandra Wachter is one who has argued that big data and AI means that current protections are inadequate and has argued that we need a right to reasonable inferences to protect us from unjust and potentially life changing conclusions about us being drawn.
- The use of data is inherently linked with profiling, where decisions are made about groups of people.
Even if a person were theoretically able to avoid having any information about them on record they would still be impacted by data held on other people judged to be like them. This is why the evidence that algorithms include inherent bias is so important – because it means that decisions about our lives are being made on the basis of harmful stereotypes.
“Behaviour described out of context is dangerously ambiguous.”
Emeritus professor Eileen Munro
No to Schools Bill campaign
Data kept on unsubstantiated accusations
Unsubstantiated allegations – or even those proven to be false – stay on records and can cause harm.
When a family is investigated the fact that they are ‘known to social services’ stays on their records, even when there was not found to be any problem with treatment of the child. The lack of evidence of harm is not seen as innocence. A ‘red flag’ comes up to notify professionals that children are ‘known’. Children and their families report that this is stigmatising and that they receive different, and often negative, treatment from professionals.
When a child is stopped and searched by the police, even when there was nothing found, that it happened is kept on records and can act as a flag making it more likely for future stop and searches.
When referrals are made under PREVENT these are kept.
PREVENT
“…even if a referral is not progressed, that is not the end for those affected: those [PREVENT] referrals remain in police, local authority and government systems in line with retention periods for genuine safeguarding issues and protocols determined for criminal issues. In fact, most people do not realise that referring an individual places that individual’s data directly with the security services as well as potentially onto local authority and multi-agency partners.”
Children’s services
“…mothers who had undertaken SARs knew that records held about them were inaccurate, contained false information and included unsubstantiated allegations that had been proven incorrect. Indeed, in some instances, professionals involved were struck off from professional practice as a result. Mothers found it deeply distressing that despite this, records about them would not be amended and the stigma of being investigated remained.”
Sarah Gorin, Professors Rosalind Edwards and Val Gillies and Dr Helene Vannier Ducasse
‘Seen’ through Records: Parents’ Access to Children’s Social Care Records in an Age of Increasing Datafication (2024)
Victim or target?
Within what is increasingly a punitive child protection and policing system being identified as a ‘victim’ brought many of the same negative outcomes as being identified as a perpetrator.
Research by Williams, Wroe and Koch found that the impact on Black boys of being included on safeguarding multi-agency databases as at risk of ‘county lines’ or ‘modern slavery’ led to the same negative outcome as that of being included on databases purportedly identifying ‘gang nominals’. Research by Amnesty found that the Metropolitan Police’s database of ‘gang’ members included victims of violence and those not suspected of any crime.
“…it seemed that being profiled by multi-agency safeguarding partners as being ‘county lines’-affiliated could have the same troubling consequences for the majority of Black boys and young men who were profiled in this way.”
Professor Insa Koch and Drs Patrick Williams and Lauren Wroe
Institute of Race Relations: ‘County lines’: racism, safeguarding and statecraft in Britain (2023)
“Data sharing between the police, housing associations, schools, job centres, the criminal justice system and the Home Office appears to lack safeguards; there is therefore a risk that these services will discriminate against already marginalised young people, with disproportionate impact on black boys and young men.”
Amnesty International
Trapped in the Gangs Matrix (2018)
(Mis)shapes society
What we choose to measure, and what we can measure, are important. A significant danger of data is that it furthers us down the neo-liberal route of focusing on the struggling individual rather than the wider conditions which cause them to struggle.
The demand for data can also be a distraction. To demand to know precisely how an injustice causes damage results in delay and confusion when our energy should be in demanding the end to the injustice itself.
“…we run the risk of measuring the impact of school absences but not the impact of school cuts. In other words, these systems, in their emphasis on correlation over causation, can individualise social problems by directing attention away from structural causes of social problems.”
Professor Lina Dencik, Dr Arne Hintz, Assistant Professor Joanna Redden & Harry Warne
Data Justice Lab: Data Scores as Governance: Investigating uses of citizen scoring in public services (2018)