Abstract
Family services in the UK are becoming increasingly reliant on technologies which merge and analyse personal data trails to inform policy and practice. Complex AI enabled tools are now regularly used to monitor and profile households, allocate resources, risk assess and target family interventions. This paper explores the impact of this technological transformation on parents by centring their views and experiences. Drawing on a mixed methods study we demonstrate how digital infrastructures are re-ordering family–state boundaries, bypassing the knowledge and consent of those impacted. The punitive, inflexible and in some cases deeply harmful consequences of data led practice for children and families are highlighted. We show how such negative encounters are cultivating mistrust and leading parents to actively avoid services seen as likely to link and profile personal data
Introduction
The landscape of family services has changed dramatically in the UK over the last 15 years. Successive governments have slashed welfare funding and service provision to invest instead in ‘data led’ mechanisms to monitor families and target interventions. This has been characterised by an increased management of resources and interventions through online portals and AI enabled dashboards. Digital sources of administrative data on children and families (e.g., health, education, social care, housing and criminal justice datasets) are now automatically stored, linked together and are in some cases pooled with commercially acquired information to profile household, assess risk and guide professional practice (Edwards et al., 2022; Gillies et al., 2025, 2021). While the economic struggles of disadvantaged families have for centuries been documented in administrative records, casefiles and other public documents, this analogue form of governance was necessarily self-limiting in its capacity and scope, circumscribed by a reliance on paper-based materials which had to be physically stored, located and accessed to inform decision making. The new digital technologies are transforming the capacities and ambitions of the state constituted by what have been loosely termed algorithmic governance techniques (Gritsenko and Wood, 2022).
Given the major implications for citizen state relations, the widespread adoption of data manipulation and machine learning tools across UK infrastructure has received relatively little sustained attention. Sweeping technological modifications have been positioned as merely structural upgrades in public sector data management, downplaying the impact on everyday lives and relationships. Governments and their agencies justify substantial investments of public money in such systems though claims to greater efficiency and the potential for future cost savings, presenting their adoption as self-evidently beneficial. However, as Dencik (2022) demonstrate, the turn to algorithmic governance has underexamined consequences for power relations, democracy and social justice. Accelerating deployment of data tools and techniques by local authorities, police and border forces, health, education and welfare agencies has been characterised by an absence of information about the impacts on those who are subject to them, particularly children and families (Redden et al., 2020).
In this paper, we highlight the risks associated with such technological restructuring and point to the unintended consequences they generate. Centring the perspectives of parents, we explore how the increased surveillance and targeting capacities of digital governance tools are understood, experienced and resisted. Our research presents clear evidence that families may be avoiding agencies that are perceived to link and analyse personal data. We begin by outlining the digital transformations that have reshaped child and family services, detailing how parents’ relationships with practitioners are now routinely mediated and shaped by dashboards, data linkage and algorithmic profiling. We then draw on survey and interview data to examine parents’ views on how governmental institutions and related agencies are using their personal information to guide decision making. Our findings point to a serious disparity between parents’ expectations of privacy, transparency and support, and the often punitive digital driven interventions many now receive. As will be demonstrated, fear and uncertainty about how personal data is being recorded, linked and used is likely to prevent increasing numbers of families from seeking crucial support.
Family services and the digital revolution
In 2017, the then Conservative Government set out a plan to ‘transform the relationship between citizen and state’ by drawing on the tools, techniques, technology and approaches of the internet age’ (Cabinet Office 2017). The Government Transformation Strategy (GTS) styled itself as ‘the most ambitious programme of change of any government anywhere in the world’, promising the digitisation of the state and the use of data as an enabler of services delivering cost savings and greater efficiencies. The GTS heralded an extensive reorganisation of government infrastructure and notably a re-engineering of welfare state delivery through greater procurement and contracting out to commercial digital platforms and expertise. Strategic goals have centred on the importance of easier and faster data sharing across services and increased utilization of advanced technology with the intention of solving ingrained social problems (Edwards et al., 2022; Gillies and Edwards, 2024; MHCLG 2021). In particular, the collection and linkage of digital sources of data on families has been pursued in an effort to target interventions and resources through algorithmically driven processing.
The first major Government service to become internet-only was Universal Credit, the social security benefit system supporting the poorest households in the country. The digitisation of this service generated considerable debate and concern with many pointing to the marginalising impact of making access to entitlements contingent on a claimant's ability to navigate complex automated online systems (Alston, 2018; CPAG 2019; Griffiths, 2021; Meers, 2020). Discussions have tended to focus on the benefits of streamlined government services versus the risks of excluding those lacking the skills and resources to engage online (Harvey et al., 2023; Wilson-Menzfeld et al., 2024). Broader objections have also been raised about the potential for such reforms to undermine the founding principles of the welfare state through the imposition of self-service, customer-centric ideologies that prioritize individualised market-based logics at the expense of broader values around inclusion, impartiality and justice (Dencik, 2022; Zamani and Rousaki, 2024).The more immediate and practical impacts of this technological transformation of welfare structures for children and families, however, have received considerably less attention.
Digitisation of administrative records works to intensify and extend the regulatory powers of the state, pushing past settled boundaries of liberal democracy. Families are increasingly subject to ‘datafication’, meaning their everyday lives and practices are translated into digital data points to analyse, monitor and make predictions about their lives (Mayer-Schonberger and Cukier, 2013).While commonly deployed to market commercial products and services, datafication and algorithmic targeting techniques are now sold to governments as ‘policy solutions’ (Gillies and Edwards, 2024). Technology companies are mainstreaming advances in computer science to upscale, amalgamate and extend government accumulation of information on families far beyond paper files and even the simple, fixed category digital material previously held in discrete, numerical, searchable databases. Unstructured data like telephone conversations, images, videos, web based transactional and interactional trails can now be stored in their raw, unprocessed state and quickly configured alongside categorised data to surveil, target and shape family interventions. User friendly digital dashboards allow state administrators and professionals multiple visualizations of targeted families, enabling interface shifts from broad geolocational patterns to ‘holistic’ profiles of individual household members which highlight and flag perceived risky characteristics, behaviours and relationships. As we discuss below, these digital insights can then be used for a variety of different purposes including identifying and tracking low-income households, assessing the likelihood of benefit fraud as well as predicting possible future incidence of social harms like child abuse, neglect, exploitation, anti-social behaviour and domestic violence.
Central to the success of this model is the extensive data profiling of families through the mass collection and analysis of service records, case notes and database entries as well as privately acquired profiling data (for example see https://www.experian.co.uk/blogs/latest-thinking/marketing-solutions/cost-of-living-support-lambeth-council/). To this end local Government authorities have been supported by central Government towards investment in new digital infrastructures with the capacity to amass, merge, and analyse extensive data on families, individual household members and their associates. Some, like Bristol and Essex County Council have collaborated with their local police forces to establish vast datahubs pooling together information from a variety of agencies to profile and target vulnerable families. Other local authorities have contracted out to the commercial sector, choosing from a wide selection of tools and services that can accessed through the central government's ‘digital marketplace’, a procurement hub set up to promote access to such data services. Companies like Sentinel Partners, Xantura and Policy in Practice offer products and services specifically tailored towards family governance and intervention These range from the design of inhouse portals, dashboards and data manipulation tools to the collection and storage of data in ‘warehouses’ or merged ‘data lakes’ for the development of AI driven prediction tools. (Gillies et al., 2021; Gillies et al., 2025)
Products currently in use include Low Income Family Trackers (LIFTs) which pull together extensive data on poor households into a structured database, allowing local authorities to identify and risk assess marginalised households (for details see https://policyinpractice.co.uk/policy-dashboard/). Others offer more sophisticated real time monitoring capacities (often termed ‘dataveillance’). For example, the data management company Sentinel Partners supply a significant number of English local authorities with complex data tools designed to monitor families to predict and prevent problems, promising ‘a fully joined-up view of all that is known and all that has gone before’ (for details see https://www.sentinelpartners.co.uk/single-view). Xantura offers a similar product range to local authorities but admit to combining administrative records with data scraped from the internet, claiming this enables identification of a wide range of family problems, even anger issues, dangerous dogs, and a history of promiscuous or unsafe sex (EY and Xantura Covid-19 Analysis, Project Briefing cited in Big Brother Watch 2021).
Attempts to operationalise data as a tool of governance are far from confined the UK, with similar automated technologies in use across the world(e.g., Eubanks, 2017; Jørgensen, 2021; Keddell, 2019; Glaberson, 2019).This technocratic turn constitutes what van Veenstra and Kotterink (2017) have termed ‘data driven policy making’, extending digital centred governance far into traditional policy realms of health, education, policing and immigration control. Proponents of such systems present them as common-sense applications of scientific progress, solving engrained social problems while simultaneously saving public money (eg. Tony Blair Foundation for Global Change, 2024). But this simplistic vision glosses over significant and growing concerns about the accuracy and impacts of such technology on targeted populations. Data profiling and predictive analytics have been found to operate with high error rates particularly when applied to vulnerable populations (Narayanan and Kapoor, 2024). Operational systems have also consistently demonstrated alarming levels of bias, supporting and compounding discriminatory decision making (Benjamin, 2019; Crawford, 2021). For example, close analysis of a family screening tool used in Allegheny County, Pennsylvania to predict children at risk of maltreatment revealed dubious claims to accuracy and clear evidence of operational bias against black families (Eubanks, 2017; Samant et al., 2023; Wang et al., 2022).
The increasing use of dataveillance and AI automation as tools of governance has also generated deep and widespread unease about the impacts on privacy, freedom and democracy (Schippers, 2020). A marked lack of transparency and accountability have supported the rapid introduction and administration of these tools. There are no central registers of public data practices and legislative changes have encouraged local authorities and their agencies to bypass data protection requirements relating to consent (Gillies et al., 2025). Such exemptions remain a legal grey area (see Children's Commissioner 2022) but forthcoming legislation is likely to further weaken data protection rights (see Chaggar, 2025). Extensive reliance on commercial companies and their technologies works to further mask knowledge of and liability for public data systems, while introducing additional risks in relation to data privacy and security (Cohen, 2019). As a result, complex data driven technologies, and the companies who provide them, exert considerable power while operating largely beyond the liberal public sphere as new centres of sovereignty (Davies, 2020).
Research has shown how a series of high-profile data breaches, cyber-attacks, data fraud and misuse of data has prompted greater public hesitancy and fear around the processing and sharing of personal information, both in the UK and the US (Ghafur et al., 2020). Yet there have been few attempts by governmental institutions to either inform the public about technological changes in governmental practices or to engage with any objections. Thus, powerful data tools have been quickly and quietly introduced across the public sector, embedding AI enabled digital infrastructure without explanation or deliberation (Edwards et al., 2024a; Gillies et al., 2021; Gillies et al., 2025). While institutions are buying into the transformative capacities of such systems with public funds, they are selling them to service users as inconsequential administrative adjustments. Public consultation about the sharing and linkage of administrative records has focused largely on anonymised data for research purposes (e.g., Moody and Lugg, 2017; NatCen, 2018), while policy discussions about the ethical challenges of tracking and profiling families are framed by assumptions that such practices are benevolent, effective and necessary (Edwards et al., 2024b).This has left a significant gap in knowledge about how datafied populations perceive and manage these technological shifts in governmentality in a context where family has traditionally been positioned as a bedrock of privacy and freedom from the encroaching powers of the state (Holvast, 2007).
Research framework
Our research centres the views and experiences of those who are most likely to be tracked and targeted by public data tools, raising important questions about the democratic mandate for such practices and the potential for unintended consequences. Drawing on a mixed method study with parents of dependent children we consider norms and values in relation to governmental uses of data linkage and analytics. More specifically, this paper is based on four key stages of our project: Freedom of Information requests/documentary research, a representative national survey, focus group discussions and individual interviews.
In an attempt to determine the data tools and practices UK local authorities are utilising in their work with children and families, Freedom of Information (FoI) requests were made to all local authorities in the UK. Intensive documentary research was also conducted to identify funding streams, public–private partnerships, reports, presentations and corporate case studies relating to the use of data tools in the governance of families in the UK. Widespread use of exception clauses and other means of avoiding disclosure significantly reduced the effectiveness of FoI as an investigatory method but substantial evidence of the commissioning and operationalisation of systems which track, merge and profile data on families was established through other documentary research (Gillies et al., 2021).
This scoping of institutional data practices fed into a UK wide survey of parents, commissioned from NatCen as part of a probability-based panel. This survey was designed to be representative of parents across the UK, to gain an understanding about what parents deem to be acceptable or unacceptable in relation to data linkage and analytics. The NatCen panel is recruited from the British Social Attitudes survey, a high-quality random probability-based survey conducted online and over the phone. It employed a sequential mixed mode fieldwork design, and weights for non-response. Randomisation of statement questions and flipping of answer option order was used to counter mode effects. The sample consisted of 843 parents with a range of demographic characteristics (see appendix), of whom 57 per cent were mothers and 43 per cent fathers. 44 per cent of the parents were in managerial and professional occupations and 50 per cent were educated to degree level or above, while 38.5 percent were in lower semi-routine and routine occupations and 10 per cent had no qualifications. Ethnically, 73 per cent of parents were White British, with 19 per cent of our sample from minority ethnic groups, five per cent of whom were Black.
Online focus group discussions were then held with homogenous sets of parents, with participants in each group sharing an element of the same social location. The choice of characteristics for each group discussions drew on findings from the probability-based survey and included parents with professional occupations, home based mothers, Black mothers, lone mothers and parents with disabled children. Topics covered generalised assessments of justifications for and oppositions to data linkage and analytics. Parents were recruited through social media and via child and family focused organisations. A total of nine focus groups were convened, comprising an average of four parents in each, involving 36 mothers and fathers overall. A majority of the participants were white, though five were Black, two were Asian and two were Mixed race. Other than the group of parents of children with disabilities, the majority in the focus groups had no contact with child or family-based interventions beyond everyday universal provisions.
The fourth stage of the research involved individual interviews with parents of dependent children who used or were subject to family support or intervention services. Parents were recruited using an advertisement about the research that was circulated widely on the internet via voluntary sector organisations and family support services, through social media platforms such as Facebook and Twitter, and snowballing via word-of-mouth. All interviews, bar one, were conducted using Zoom or by telephone. All had been in contact with some form of family support service, although the length and type of this contact was variable, ranging from parents who had accessed parenting classes or services at local family centres to those who had extensive involvement with Children's social care, some of whom had experienced removal of children from their care. A total of 23 individual interviews were conducted with 20 mothers and 3 fathers from across England, Scotland and Wales. Approximately half were homeowners receiving average or above incomes. Participants were predominantly White but included one Black and two Asian parents. Both focus group and individual interviews were analysed using inductive code and theme development. Pseudonyms are used throughout this paper. Ethical approval for the survey, focus group and individual interview data was given by the University of Southampton ERGO 56997 on 12.2.21. Data generated by the study is available on registration with the UK Data Archive.
Parents views on institutional data linkage
Our findings point to broad uncertainty and unease felt by parents about public data practices. A large majority of our survey sample (82%) do not feel that parents know or understand how agencies are using their information. This reflects not just a broad lack of consultation with families in designing, commissioning and operationalising institutional data platforms, but a more general disregard of the public interest in how personal information is handled and processed by governmental institutions. Even concerted efforts to uncover details about particular systems and practices are commonly thwarted. As we discovered in our attempt to map widely-used data systems through FoIs, local authorities either could not or would not provide full details about the operational systems they use to link and profile data on children and families (Gillies et al., 2021).This silence has left many parents completely in the dark about technological changes that may substantially shape and alter their interactions with institutions and agencies. While our survey suggests most parents are aware that administrative records are collected and digitally stored (72%) only just over half of knew that digital administrative records from different sources can be joined together to find out more about individual families (53%) (Edwards et al., 2022).
Unfamiliarity with operational data procedures and capacities also emerged as a feature of our focus group discussions, with participants seeking clarifications and definitions from the researchers about how their data might be processed. Widespread lack of awareness of common governmental data procedures sat alongside a belief that there should be greater openness. The survey revealed a clear consensus that Governments should publicise their joining together of administrative records about families and explain how this information will be used, with 81% of participants expressing this view. Strong values were also evident in relation to principles of consent. Most parents believe that agencies should explicitly acquire permission before data is linked. Only 14% of surveyed parents thought it was acceptable for administrative records to be joined together without consent.
Consent also emerged in the focus group discussions as an important guiding principle of public data management. Indeed, many participants believed that governmental access to their data was restricted without their permission and were surprised to find this was not the case. For example, Evie in the home-based mothers focus group initially assumed that prior consent was a necessary condition of data sharing, limiting government access to information on families. Her statement to this effect quicky gave way to doubt in the context of the focus group, but the discussion generated universal agreement that consent was an individual right that agencies should respect. Evie: I mean, obviously there's got to be some kind of level of consent involved before an agency share (pause) I mean, yeah, I guess that's the question. I mean, with the research you’re doing, will, do agencies need to seek first our consent before they share this information? Or is it going to be able to be shared without our consent…… Delilah: They should have our consent, because it should be our rights as well…. Evie: I agree. Jasmine: I agree…
Concern over the involvement of the commercial sector in the technical manipulation of public administrative data marked another point of consensus. Only 39% of the survey participants felt it was acceptable for local authorities to contract out these services. In focus group interviews this mistrust of industry partnerships was even more pronounced, with concerns centring on a perceived lack of corporate oversight and the potential ulterior motives governing data companies. Participants commonly drew on examples of unprincipled behaviour of large tech companies to support their misgivings, as Hallie's comment in a lone mothers focus group illustrates: I’d be looking for kind of the motivation of these companies because we’ve had like the Facebook scandal and like Google scandal where they’ve sold this information and where you get your, you know, you look on for a dress and then you get 54 messages come up. A moral and ethical private business that fundamentally protects individuals’ data at the expense of shareholders’ returns doesn’t exist, because they would be out of business, by definition. So, I, on a personal level, am wary.
Data led targeting: Implications for families
While focus group parents were often shocked to discover that statutory agencies might link, profile and share digital information about them, participants in the service users’ sample had much greater familiarity with agencies’ data practices. Their awareness that personal data can be tracked and used to assess risk was often painfully acquired through challenging, and in some cases traumatic encounters with professionals. Trust in the veracity and efficiency of these data-centred practices was generally low among these parents, shaped by experiences of error-strewn record keeping and miscommunication. Many also complained about highly subjective and pejorative judgements made about their parenting practices by individual workers, leading to false allegations that they were harming their children (Gorin et al., 2023). Notably, half of all parents in our sample of service users had at some point been subject to a child protection referral, with information sharing as a factor, though none were undergoing any resulting statutory interventions at the time of the interviews.
This high number likely reflects a marked increase in UK parents being referred and formally investigated by social services. As as Jay et al. (2025) have demonstrated, one in three children born in England since 2011 have received referrals to child protection agencies, while formal investigations finding no harm have tripled (Bilson and Drayak, 2020; Bilson and Hunter Munro, 2019). This is, in part, due to increased referrals from schools, evidenced by a 50% rise in teachers referring to social services since 2014 (Walker, 2022). However, among our interview sample, most of the contact with social workers occurred after an interaction with health care services because of linked data profiles. For example, one mother had received an unannounced visit from a child protection social worker after she took her daughter to hospital on finding the baby's own nail had cut the inside of her mouth. In five other cases mothers had been accused of harming their children after seeking medical care for them. Two had been suspected of physically abusing their children before an undiagnosed medical condition was found to account for the symptoms. The three other mothers had children with complex health needs and had been accused of either fabricating or exacerbating these conditions. In two cases children had been removed and temporarily placed in the care of the local authority, before being returned by the court. In both these cases institutional malpractice was established, though not before one child was sexually abused while in local authority care.
Parents across the sample of service users were concerned by how information about their family had been recorded and shared. Some emphasised their alarm at finding long past experiences brought to light to inform and guide risk assessments and interventions. One mother subject to a child protection investigation described how a social worker ‘dredged up’ her troubled teenage experiences from 25 years ago, while another felt her experiences of being in care as a child had been used to position her as a ‘lesser person’. Many of these parents worried about the power government agencies and professionals hold over families through the control and management of their data. In particular, the parents who received social care contact were left deeply distressed by their experiences and described feeling frightened and powerless. Most remained acutely aware that even in cases where allegations were unsubstantiated, risk alerts remained on data profiles and summaries, shaping and potentially prejudicing future contact with services. Lenna's daughter spent 18 months in local authority care before a catalogue of malpractice and professional failings were established. Yet, as she explained, the original child protection concerns remain on her record while the mistakes that escalated them are invisible: The absolute awful badness of it, to have that [false allegation] linked and perpetuated and replicated on different systems and for people who I've not even met yet, you know, like a new CAMHS nurse might come and say oh I'll just find out a bit more about this family before I see them the first time and that's what she's reading…..She's not seeing the report….She's seeing red flags, red flags, child protection register.
Because parents are not routinely able to view their own data profiles, the identification and challenging of erroneous entries, assumptions and misinterpretations is difficult. Exercising data subject rights was experienced as a frustrating and laborious process. Five parents in our sample had submitted a Subject Access Request to better understand what had happened to them or to hold professionals to account, but all described being actively ‘blocked ‘or having their requests denied by the institutions they applied to. The more persistent parents managed to obtain some heavily redacted copies of service files, but none were able to retrieve a full record of the data held about them (see Gorin et al., 2023 for extended discussion). Moreover, the processing, evaluations and data visualisations generated by the agency software tools remain hidden, substantially tipping the balance of power against parents. The abstraction and reconfiguration of data points into classifications, scores and ‘red flags’ strips out social and cultural meanings to present a simple and actionable picture to professionals. But lack of context often led to skewed interpretations and inappropriately heightened impressions of risk. For example, Diane described how a record of missed medical appointments triggered a section 47 child protection investigation. The fact the appointments had been cancelled by the hospital had not been documented. In Tracey's case, a social work report noted that her son had been found outside with no shoes and his coat open, with no mention that he had been playing on a trampoline at the time.
Mistrust, avoidance and fear
As we have highlighted, the impacts of data linkage and profiling can be profound and unjust, but our research also points towards broader consequences for public trust and wellbeing. Our survey suggests parents have relatively low levels of confidence in services to strategically use family records. Only around half of all participants felt comfortable with children's social work teams, local councils, education services, early years services, police and criminal justice, or immigration routinely accessing or merging administrative data profiles to target support for families. Among more marginalised social groups, trust in the data linkage practices of these individual services was considerably lower, particularly among Black parents, those on low incomes, young parents, and large families (Edwards et al., 2022; Edwards et al., 2024b).
Moreover, Black participants in the survey were much more likely than other ethnicities to feel that data linkage leads to greater discrimination and prejudicial outcomes. This concern was also raised and elaborated on in-depth in the focus group discussions. Black mothers in particular were acutely aware of the dangers of hidden biases and the assumptions that might be projected on to their family's data profiles, with many describing previous discriminatory or negative encounters with services. For example, Evie, a Black participant in the at home mothers focus group, clarified her opposition by explaining how she had discovered an inaccurate, hurtful and discriminatory note on her son's medical records. She had been unable to get the note removed, despite receiving a written apology from her medical practice. She felt this experience exemplified the dangers of linking data across services. In both Black mothers’ focus groups there was a shared sense that technological developments in data targeting had been characterised by dishonesty and lack of transparency. Eberia felt she had been ‘lied to’ about her health records being private and worried that any reliance on statistically generated categorisations was a slippery slope to eugenics.
Some parents across the focus groups and services users’ interviews were supportive of the broad principle of using data linkage and analytics to direct early help to families, but widespread scepticism was expressed that this was a core motive or likely outcome of such practices. Parents pointed to the existing over-demand and under-resourcing of services to question how such an approach could work without a substantial injection of funds, with many describing themselves as ‘cynical’. Parents from the service users sample were particularly unconvinced by the claim that data targeting would increase access to support: If we're not going to put the resources into supporting these things properly, you can have all the information you want in the world, but if it's not backed up by a really holistic service that's fully funded, then you could be opening wounds for people or creating issues for people and then not supporting them. (Jade – service user) People are asking already for help but they don’t meet thresholds, they don’t meet criteria. So how are they going to do it, make it better sharing information? (Naomi – service user) It feels like there's sneakiness is going on which a lot of us, well I assume, feel like that anyway generally in society, you know, with these kind of like, you know, services. My GP records would show I’ve been diagnosed and medicated for depression several times in the past. It's never affected my children's welfare or my employment. But now I feel like I’m on a final warning, you know, because that's strike one. If I lose my job now or like I get – like the police, actually had to move my car recently when I broke down. So I’ve got a crime reference. So is that strike two? (Jo – lone mothers focus group) I hear from a lot of parents, as a parent peer support and as a parent of a kid..‘I don’t want to take my child to the doctor, because then they think this’…… There's dozens of forums online, where parents are telling other parents, ‘Don’t ask for help because it will be used against you.’ That's a big risk, we could end up with children who are actually getting hurt because their parents are too scared that they might lose them if they ask for help.
Conclusion
Our research points to a huge and worrying disconnect between the values and expectations of parents and institutional data practices. As we have shown local governments and related services across the UK are increasingly utilising complex digital tools that link and analyse data on families. Data lakes, predictive modelling and automated dashboards have been quietly incorporated into children and families’ services, leaving parents largely unaware of the extent to which their information is being processed to classify, risk assess and target intervention strategies. Despite cutting across longstanding norms relating to privacy and individual rights, remarkably little consideration has been given to the views, perspectives and experiences of families in the design, commissioning or operationalisation of such tools. Moreover, as our research established, information about these systems and their classifications of families are remarkedly hard to come by (Gillies et al., 2021; Gillies et al., 2025).
The opaque nature of these digital practices and lack of any broader public debate led some parents participating in our research to voice shock and strong disapproval. The absence of consent protocols mediating agencies’ use of their data was widely disavowed, as was the involvement of private companies in managing personal information. While most participants in our focus groups had limited knowledge about agency data practices, parents who were regular service users were more likely to recognise their significance and to have had extremely negative personal experiences. A high number of these interviewees gave harrowing accounts of being investigated, falsely accused and stigmatised on the basis of long past incidents, data errors or misinterpretations. These cases provide a disturbing insight into the real human and social costs of increasing reliance on digital tools for welfare delivery, particularly in relation to child protection services. As we have shown, the fear and disempowerment such practices generate are eroding trust and driving parents away from crucial services.
The techno-solutionist agenda propelling the introduction of data tools into welfare services is remarkably detached from moral arguments and justifications in relation to family lives. Our research highlights the potential for anger and suspicion of institutional data practices to grow with increased exposure, as the implications for individual rights, civil liberties and democratic values crystalise among the wider population. That such a major reorganisation of welfare services has proceeded at pace without meaningful public consultation or even adequate disclosure, suggests that Governmental organisations are gambling public trust against promises of future cost savings and greater efficiency. This stealth pursuit is particularly high stakes in a context where public trust in governmental organisations have reached new lows (Montagu and Maplethorpe, 2024).A s we have demonstrated, data-led practices are amplifying mistrust and apprehension with the effect that those most in need of crucial family services are actively avoiding them.
Meanwhile, the odds are stacking up against the longer-term effectiveness of public data ‘solutions’, even without accounting for the costly and damaging consequences of service avoidance. Their value for money remains unevidenced, while empirical evaluations of their predictive modelling capacities show very poor reliability in real world situations (Salganik et al., 2020). Levels of UK state spending on digital procurement and outsourcing have soared in recent times, leading the director of the National Audit Office to quietly upbraid government departments for their ‘unrealistic ambitions’ and reliance on ‘untested technology’ (Clark 2022). While Governments are placing ever greater faith in the power of digital tools to solve complex family problems, data scientists are increasingly acknowledging the limits of these technologies and denouncing the commercialised hype that surround them (Narayanan and Kapoor, 2024).Critics point to a growing list of calamitous data driven policy tools, many causing grievous harm to families (see Narayanan and Kapoor, 2024; Wang et al., 2022). These include a welfare algorithm introduced in the Netherlands that wrongly accused at least 35,000 parents of childcare benefit fraud. The accused families, predominantly from racialised minorities, were plunged into severe debt and protracted legal proceedings. This resulted in forced evictions, bankruptcy, the removal of over 2000 children from their parents’ custody, and ultimately the downfall of the Dutch Government in 2021 (Arts and Van Den Berg, 2025). In 2025 similarly flawed technology was trialled in the UK and led to the wrongful suspension of Child Benefit payments to thousands of families (Butterly and O’Carroll, 2025). The regular unfolding of such scandals in the media seem likely to intensify parents’ fears and suspicion and could even precipitate a rapid and widespread disengagement from agencies.
In a context of systemic underfunding of regional governments, health, education and related services the appeal of a technological cure for engrained social ills is obvious. However, there is little reason to believe that complex AI enabled data tools will fulfil their inflated promises and much evidence to suggest they could cause significant damage to children and family's lives. As our research suggests, intrusive, dysfunctional and widely distrusted data practices may have a costly momentum of their own. Families who are harmed rather than helped by state services are likely to disengage altogether, deepening their marginalisation and compounding the very problems digital ‘solutions’ are designed to fix.
Footnotes
The research on which this article is based is funded by the UKRI Economic and Social Research Council under grant number ES/T001623/1.
.
Author biographies
Recent publications include an article in Journal of Social Policy Research and a chapter in The Conservative Governments and Social Policy (Bristol University Press, 2024).
Inclusion of Children and Young People with Special Educational Needs and Disabilities in Schools – How Can Local Areas Support Schools?