Socialinė teorija, empirija, politika ir praktika ISSN 1648-2425 eISSN 2345-0266
2025, vol. 31, pp. 119–132 DOI: https://doi.org/10.15388/STEPP.2025.31.8
Goetz Ottmann
Federation University Australia, Australia
E-mail: g.ottmann@federation.edu.au
https://orcid.org/0000-0002-3568-2166
https://ror.org/05qbzwv83
Mara Sanfelici
University of Milano-Bicocca, Italy
E-mail: mara.sanfelici@unimib.it
https://orcid.org/0000-0001-6588-5338
https://ror.org/01ynf4891
Carolyn Noble
Australian College of Applied Psychology, Australia
E-mail: Carolyn.Noble@acap.edu.au
https://orcid.org/0000-0003-3007-947X
https://ror.org/027wzg564
Ines Martinez Herrero
Universidad Nacional de Educación a Distancia, Spain
E-mail: m.ines.martinez@der.uned.es
https://orcid.org/0000-0002-7743-2771
https://ror.org/02msb5n36
Caroline Bald
The Open University, United Kingdom
E-mail: caroline.bald@open.ac.uk
https://orcid.org/0000-0001-7324-2738
https://ror.org/05mzfcs16
Eglė Šumskienė
Vilnius University, Lithuania
E-mail: egle.sumskiene@fsf.vu.lt/
https://orcid.org/0000-0002-8645-5748
https://ror.org/03nadee84
Melisa Campana
Universidad Complutense de Madrid, Spain
E-mail: melcampa@ucm.es
https://orcid.org/0000-0003-3988-8273
https://ror.org/02p0gd045
Andres Arias Astray
Universidad Complutense de Madrid, Spain
E-mail: aariasas@ucm.es
https://orcid.org/0000-0001-8614-0714
https://ror.org/02p0gd045
Renata Nunes
Universidad Complutense de Madrid, Spain
E-mail: renunes@ucm.es
https://orcid.org/0000-0001-9781-4067
https://ror.org/02p0gd045
Abstract. Technologies of governance, surveillance, and control have evolved over the course of human history and are transforming social work as they are deployed across a plurality of government, private sector organisations, and social media networks. In this article, we explore key issues associated with automated algorithmic governance in relation to social work education and practice, with particular attention to the implications for professional and institutional resilience. Focusing on the datafication of welfare governance, we trace its historical antecedents and applications, and provide a brief overview of digital automation and data-driven governance in Lithuania, Italy, the UK, and Spain. The aim is to identify issues that warrant further attention in social work curriculum design, particularly regarding the governance of state-financed social support services, and to argue that strengthening resilience requires social work to become more aware, technologically literate, and politically engaged in order to advocate for and support service users within increasingly digitally enabled and networked systems of control.
Keywords: Resilience, digitalisation, algorithmic governance, surveillance, social work.
Received: 2025 12 12. Accepted: 2026 02 09.
Copyright © 2025 Goetz Ottmann, Mara Sanfelici, Carolyn Noble, Ines Martinez Herrero, Caroline Bald, Eglė Šumskienė, Melisa Campana, Andres Arias Astray, Renata Nunes. Published by Vilnius University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Across welfare states, automated algorithmic decision-making (AADM) is increasingly being integrated into the governance of social services (Juell-Skielse et al., 2022; Weatherall et al., 2024). These systems are used to classify risk, determine eligibility, allocate resources, and monitor compliance, thereby reshaping how decisions are made within welfare administrations. (Mann, 2020; Ottmann, 2020). Indeed, surveillance has become a key aspect of bureaucratic administration, and the provision of human services is largely bound up with the collection, analysis, and deployment of societal and personal information. While a growing body of scholarship has examined the ethical, legal, and technical implications of algorithmic governance, considerably less attention has been devoted to understanding how these developments interact with the historical trajectories of welfare systems and reshape professional resilience within social work across different national contexts.
This article analyses the role of AADM in expanding surveillance-based governance and to assess its ethical implications for social work within a broader historical context. Taxonomies, decision trees, and structured decision-making frameworks are widely used to consolidate professional identities, practice, and to sharpen professional boundaries. It is assumed that by reducing and replacing subjective discretionary decision-making with objective standardised procedures will increase the quality of social work practice. In doing so, AADM tools, such as the Australian Integrated Assessment Tool, often gloss over protected attributes (e.g., gender identity, sexual orientation, religion) that, in practice, translate into important differences in the name of an administratively defined inflexible procedural fairness but that contribute to outcomes that are essentially unfair and that undermine human rights. AADM shifts the ability to frame the profession away from social work peak organisations and service providers to statisticians, systems architects, software designers, and politicians.
While often presented as a means to improve efficiency, reduce errors, and enhance service delivery (Eggers et al., 2017), digital automation has frequently been implemented to control work practices, constrain demand, reduce costs, and regulate the service users’ behaviour. Emerging research highlights that such processes may dehumanise service users, undermine rights, and reshape professional roles, practices, and knowledge systems (Benjamin, 2019; Eubanks, 2018; Schuilenburg & Peeters, 2021).
The article situates AADM within its historical antecedents and examines its implications across Lithuania, Italy, the UK, and Spain, with a focus on social work education and governance. The aim is to tease out issues that must be explored in social work education focusing on the governance of state-financed social support services arguing that social work will need to become more aware, tech-savvy, and politically involved if it is to advocate for and support service users captured by digitally enabled, networked control (Monahan, 2017). Throughout this article, resilience is understood not simply as the capacity to adapt to technological change but as the ability of professionals, organisations, and welfare systems to maintain ethical judgement, democratic accountability, and human rights while responding to expanding forms of AADM. This conceptualisation shifts resilience from an individual characteristic towards a broader institutional and political capacity.
This article contributes to the emerging literature in three respects. First, it conceptualises AADM as a historically embedded form of welfare governance rather than merely a technological innovation. Second, it demonstrates how different welfare traditions shape distinct configurations of professional and institutional resilience. Third, it argues that resilience should be understood not simply as adaptation to technological change but as the capacity to critically shape the ethical and political conditions under which AADM is implemented.
AADM takes multiple forms, including risk profiling, resource allocation, compliance monitoring, and predictive analytics. The most common form of digital automation is based on relatively simple decision-tree processes (Weatherall et al., 2024). Within social work, AADM is used to identify welfare ‘fraud’ and calculate risk profiles within child protection (Eubanks, 2018; Keddell, 2019), domestic violence interventions, and criminal justice contexts. What we are currently seeing in the public service, is an explosion of such AADM systems (Weatherall et al., 2024). A recent Australian report identified 213 major systems that were either already automated or planned to be automated across the state and municipal governments of the State of New South Wales. Around one third of these systems were fully automated (Weatherall et al., 2024). Indeed, Australia (5th) and the UK (3rd) are in the top tier of the OECD’s Digital Government Index. Lithuania (14th) and Spain (15th) are around the OECD average, and Italy (19th) is situated in the mid-field (OECD, 2025).
AADM has to be viewed within its political context. It aligns well with a neo-liberal trope that seeks to privatise welfare services, cut costs, and expand surveillance capacities, thereby limiting discretionary decisions in professional practice (Abramovitz & Zelnick, 2015). AADM has the potential to limit discretionary decisions even further or eliminate human decision making altogether (Ottmann, 2026). As such, digital automation reconfigures the conditions under which professional and institutional resilience can be exercised, potentially constraining the capacity of social workers and organisations to respond flexibly, contextually, and ethically to complex human situations.
This makes it possible for sections of social work frontline work to be replaced by automated processes supplemented by subcontracted call-centre operators without human services qualifications (Ottmann, 2026). Furthermore, even in contexts where digital automation is taking more of a service user-focused trajectory, significant benefits seem to flow in the direction of consulting firms and corporate software developers that are brought in to drive digital automation rather than service users, public service organisations, or the taxpayer (Gustafsson, 2022; Juell-Skielse et al., 2022). For example, the Australian government was estimated to spend AUD 19 billion on digital transformation initiatives in 2024. At that time, it was projected that around 70% of government agencies would be using automated and AI-assisted decision-making (AADM) or Artificial Intelligence to support human decision-making by 2026 (Gartner, Inc., 2024). AADM shifts the locus of control towards a select number of managers, systems engineers, and politicians allowing for legal, democratic, and bureaucratic modes of governance to be bypassed (Ottmann, 2026). It can be used to centralise and expand coercive surveillance by creating a seamless web of digital datapoints that can be easily deployed to discipline those deemed to contravene the directives of coercive states (Ottmann, 2026). Such centralisation raises critical questions about the resilience of governance systems themselves – whether they retain the capacity for accountability, contestation, and correction, or whether they become increasingly rigid, opaque, and resistant to ethical intervention.
Given these risks, effective safeguards are essential to prevent the emergence of what Agamben (1998) conceptualises as “zones of exception”, where individuals are reduced to objects of control through opaque and unaccountable systems. For social work, this raises urgent ethical and practical challenges. Professionals must be equipped to recognise and respond to unjust, discriminatory, or oppressive automated processes.
While the form and application of digital automation differ widely, all forms involve the quantification of information. As such, they promote a shift toward quantitative epistemologies, requiring complex events and contexts to be simplified into standardised inputs for computational processing. This process risks ‘flattening’ the lives of social work service users (Henman, 2020), reducing them to data representations that may misrepresent lived experiences. The integration of large, interconnected datasets further requires the harmonisation of categories across administrative systems, often developed for different purposes. This increases the likelihood of bias, misrepresentation, and exclusion, particularly when categories fail to capture the complexity of social realities, or when they are shaped by implicit normative assumptions (Eubanks, 2018). This, in turn, raises critical questions about epistemic resilience – the capacity of social work knowledge systems to retain complexity, contextual sensitivity, and interpretive depth in the face of pressures toward standardisation and quantification.
The drive for digital automation forms part of a wider narrative disseminated by consulting firms, tech corporations, and think tanks that see AADM and AI as fixes for a ‘dilapidated’ welfare state (Eggers et al., 2017). This technocratic shift narrows the space for democratic deliberation and raises concerns about the resilience of welfare systems as arenas of political negotiation, contestation, and collective decision-making. This leads to a depoliticization of the social, foreclosing what ought to be democratic debates and processes. In this sense, automated algorithms might come to replace regulation and policy, and open the door to tech start-ups whose products are in direct contravention of human rights and their embodiment in national legislation, such as privacy and anti-discrimination laws (Hill, 2024).
The privatisation of ever greater tranches of social service provision, particularly in North America, UK, and Australia, heightens the pressure on social workers engaged by the state that are under pressure by their code of conduct to deliver services in a humane, empowering way and a political class that insists on productivity and the corporatisation and privatisation of welfare (see, for instance, Ottmann, 2026). The regimentation of decision-making and erosion of professional discretion reduce service users to quantifiable indicators used to predict and ‘nudge’ behaviour, reinforcing system-driven interventions. In such contexts, professional resilience is not only challenged but may be redirected toward compliance with system logics rather than critical engagement with them. Reliance on automated algorithms to implement coercive policies is facilitated by the fact that algorithms do not query or question decisions, they do not leave a paper trail, and are less transparent in terms of followed procedures and decisions made – as they simply reproduce and apply power and action scripts that quickly become part of a new mode of governance (Juell-Skielse et al., 2022).
AADM harbours a positivist utopianism that holds that data-driven governance and the right algorithms will help us to progress towards better forms of existence. However, a growing number of evaluations focusing on child protection, predictive policing, and re-offending risk predictions (see, for example, Eubanks, 2018) highlight that this technocratic utopia harbours inherent biases, discriminatory views, potential misuse, and autocratic tendencies leading to inappropriate or harmful outcomes. The digital automation will entrench the ideological, ethical, theoretical, and socio-political assumptions of their makers (Latour, 1992) which ultimately can undermine the rights of targeted individuals and communities. From this perspective, resilience requires not only the capacity to adapt to technological systems but the ability to critically interrogate, contest, and reshape the ideological and normative assumptions embedded within them. This development has historical precedents where algorithmic governance was used as a tool for coercive control of targeted populations and where social welfare workers were complicit in its use (Valentine, 2019).
Taken together, these perspectives suggest that algorithmic governance should be understood not merely as a technological innovation but as a transformation of the epistemological, normative, and organisational foundations of welfare systems. To explore how these dynamics unfold in practice, the following section outlines the comparative methodology used to analyse four national cases.
Guided by the conceptual framework outlined above, the comparative analysis examines how AADM is interpreted and institutionalised across different welfare systems, and how these differences shape professional and institutional resilience. Comparative case studies are well suited to analysing complex socio-political phenomena embedded in specific historical and institutional contexts while preserving contextual complexity. Rather than evaluating individual digital technologies, the study examines how digital transformation intersects with the historical development of welfare systems, professional cultures, and governance arrangements. The aim is to develop conceptual insights into the relationship between algorithmic governance and professional resilience rather than to test causal hypotheses.
Four national contexts were purposively selected: Spain, Lithuania, Italy, and the United Kingdom. These contexts capture variation in welfare traditions, histories of social work development, and approaches to digital transformation while sharing common pressures associated with the growing use of data-driven governance. Lithuania represents a post-socialist welfare system whose institutional development has been shaped by the legacy of state socialism and subsequent welfare reforms. Italy and Spain reflect Mediterranean welfare regimes characterised by decentralised governance, strong relational traditions in social work, and comparatively fragmented pathways of digitalisation. The United Kingdom has been an early adopter of digital welfare governance and provides important examples of the expansion of automated and AI-assisted decision-making in public services. Together, these national contexts enable comparison across contrasting institutional histories and governance traditions, allowing us to examine how different historical trajectories shape professional resilience in response to digital transformation.
All cases were analysed by using a common analytical framework comprising four dimensions: (1) the historical development of welfare governance; (2) the emergence and application of automated algorithmic decision-making; (3) implications for professional practice, including discretion, accountability, and relationships with service users; and (4) resilience, understood as the capacity of social work systems and professionals to critically engage with, adapt to, and challenge technological forms of governance that may undermine human rights and professional values. The analysis draws on published academic literature, policy documents, government reports, and the authors’ contextual expertise in the countries examined. The purpose is not to provide exhaustive national histories but to identify broader conceptual patterns in the relationship between AADM and resilience.
Coercive forms of surveillance-based algorithmic governance are a hallmark of totalitarian regimes. Mass-surveillance that is then analysed making use of algorithmic decision trees is a key ingredient of totalitarian governance. During the 20th century, their deployment can be traced to the use of institutional databases and the deployment of psychiatrists and social workers leading to the extermination of people with disability during National Socialist Germany to the use of health and social services records to identify and crush those deemed ‘enemies’ in Franco’s Spain. These historical examples illustrate not only the expansion of surveillance infrastructures but also the systematic erosion of institutional and professional resilience, as welfare systems became instruments of coercion rather than protection. The following cases are not intended as comprehensive national analyses, but as illustrative configurations of how resilience in digitally transforming welfare systems is shaped by different historical trajectories, institutional arrangements, and professional cultures.
In Spain, resilience is memory-based. During the Franco dictatorship (1939–1975), social assistance workers and volunteers were implicated in indoctrination and family segregation policies, including “Stolen Babies and Children” (Martínez Herrero, 2024). Individuals deemed enemies of the regime were identified, categorised, and persecuted through extensive surveillance systems. Authoritarian public services, supported by networks of ‘inspectors’, collected data from banks, land registries, and communities to dispossess targeted individuals (Cobo Romero et al., 2011). This reconfiguration shifted social work from care toward coercive governance, undermining ethical resistance. While Francoist repression relied on human surveillance, contemporary digital systems can replicate and intensify these functions. The growing use of predictive analytics in policing and governance (Barber, 2024; Eubanks, 2018) raises concerns about the resilience of welfare systems to authoritarian uses.
Social work has developed a prominent digital specialisation prior to COVID-19 (López Peláez et al., 2018), however, this trajectory has been largely techno-optimistic, with limited ethical and political critique. This gap is particularly concerning in light of a far-right shift among younger generations, including more favourable views of the Francoist past (RTVE, 2025). The convergence of rapid digitalisation and fragile democratic commitments underscores the importance of historical and professional memory in shaping ethical approaches to AI in social work.
Lithuania reflects a form of adaptive but potentially insufficiently critical resilience. The country experienced another form of totalitarian coercive governance during the 50 years it formed part of the Soviet Union (1940–1990) dubbed ‘the period of silence in social work’ (Bagdonas, 2001). After the restoration of independence in 1990, Lithuania had to reform the Soviet social security model and, as other Post-Socialist countries, transition from centrally planned, state-dominated social welfare systems to more pluralistic models (Buck-Morss, 2008; Read & Thelen, 2007). This shift involved the introduction of NGOs, private providers, and community-based services into the social welfare landscape to make social services more responsive to local needs and conditions, greater autonomy was granted to local governments. This rapid transition can also be understood as a form of institutional ‘leapfrogging’, where standards, and practices from Western welfare systems were adopted within a relatively short time frame (see, for instance, Palkova, 2024). While this enabled significant progress in aligning with human rights-based approaches and modern service delivery, it also limited the time and space for critical reflection. In the context of ongoing digital transformation, this historical trajectory may shape how new technologies are received and implemented. A comparatively high openness to innovation, combined with institutional pressures to modernise, may reduce critical scrutiny of algorithmic governance and its implications. From a resilience perspective, this raises the question of whether adaptive capacity has been sufficiently accompanied by critical and reflexive capacities necessary to anticipate and mitigate emerging risks (see, for instance, Petružytė et al., 2023).
Moreover, elements of institutionalisation, control-oriented practices, and stigmatising attitudes persist within the system; thus, digital automation and AI technologies entering the field of social work in Lithuania encounter a profession shaped by both coercive historical legacies and ongoing reform efforts (Šumskienė et al., ٢٠٢٦). While significant progress has been made over the past decades, the coexistence of past and present logics creates a complex terrain. In such contexts, resilience cannot be assumed as a stable achievement but rather as an ongoing, contested process. On the one hand, it can be strengthened through critical engagement, though, on the other – it can be weakened through the unreflective adoption of new technological forms of governance.
Italy suggests a form of relational and practice-based resilience expressed through professional resistance. Here, the processes of standardization and datafication of practice have been much slower, compared to other contexts, for different reasons. While the country has a seventy-year history of professional social work, originated after the fascist dictatorship (1922–1945) as a means for democratization, it has been positioned within a Mediterranean type of welfare regime, characterized by an historical under-investment in social security, a lack of national governance, and strong territorial differentiations, at least until recently (Pavolini & Saraceno, 2022). The delivery of social services mirrors the country’s complex multi-level governance structure. Municipalities are responsible for the provision of social services, and yet, even within a common national legislative framework, they are allowed to use different guidelines and social work assessment tools.
Recent investments in public sector digitalisation have led to the introduction of electronic information systems, many of which are developed in collaboration with private IT providers. A national Information System, supported by the World Bank, has been introduced to standardise aspects of case management, support data collection, and facilitate eligibility assessments (MLPS, n.d.; World Bank, 2023). Some processes within the system are partially automated, including the transformation of pre-assessment data into predefined intervention pathways and the monitoring of service user compliance.
The Italian case highlights that the key issue is not categorisation per se, but how categories are operationalised within practice (Benjamin, 2019). In Italy, a constructivist, relational approach (Fargion, 2013; Folgheraiter, 2007; Sanfelici, 2024) still underpins much of social work practice and seems to make these professionals more inclined to resist the standardisation of digital surveillance. That is, standardised managerial processes that are imposing categories with more rigid boundaries that can potentially hamper the process of co-construction of meaning and generate reciprocal trust with service users are regarded with suspicion. Since the digitalisation process is very recent, it is still unclear which effect this new policy of standardisation will have in the long run.
The UK illustrates a form of institutionally constrained resilience resulting from the profession’s limited engagement in shaping technological governance. Given the increased use of digital surveillance to enhance the control of social services’ users, it was not unrealistic to have hoped for social work representation when the UK hosted the AI Security Summit, resulting in the Bletchley Declaration, which underlined that AADM must be “human-centric, trustworthy and responsible” (Gov.UK, 2023b). AI as being “used for good and for all” was expected to include public services such as health – and social work was noteworthy because of being absent.
This absence is particularly striking given the UK’s historical position as a leader in social work education (1908) and the first country to introduce professional registration (2001). The country has also been at the forefront of digital innovation in social work, with a growing evidence base highlighting both the potential and risks of technological transformation (Ballantyne, 2017; Taylor, 2017). However, concerns persist that digitalisation is often framed as an inevitable progression, with insufficient attention to ethical implications and professional readiness.
The global impact of COVID-19 led to an expansion of digital social work practice and opened the door to other AI ‘advances’ such as rolling out the Meeting Co-pilot software. AI trials are framed as ‘timesaving’ or ‘professionalising’. The shift to digital social work begins in education with virtual education admissions, hitherto considered counter to ethical student recruitment (Bald, 2023).
The UK government has committed GBP 21 million to AI deployment in healthcare, while promoting efficiency through “decision-support tools” (Gov.UK, 2023a). In contrast, no comparable investment or strategic focus exists for social work. While local authorities are introducing digital systems, this occurs without central coordination or profession-wide oversight. These developments reduce opportunities for the profession to influence the institutional design of digital welfare systems, thereby constraining resilience at the governance level.
Taken together, these country cases illustrate that the integration of digital technologies into welfare systems follow a broadly similar, yet not uniform trajectory, unfolding through historically and institutionally specific configurations of resilience. While digitalisation is both unavoidable and undeniable, some contexts demonstrate adaptive capacities shaped by rapid transformation, others reveal forms of professional resistance or, conversely, a diminished capacity to critically influence technological change. These variations suggest that resilience in digitally transforming welfare states is neither given nor stable, but contingent upon the interplay between historical legacies, institutional structures, and the positioning of social work within broader governance processes.
The findings support previous research suggesting that digital technologies are never introduced into institutional vacuums but become embedded within existing governance arrangements and power relations (Benjamin, 2019; Latour, 1992; Schuilenburg & Peeters, 2021). Our comparison extends these arguments by demonstrating that historical legacies of welfare governance also shape the forms of professional resilience that emerge in response to algorithmic governance. While previous scholarship has largely examined algorithmic governance from the perspective of technological bias or administrative efficiency (Eubanks, 2018; Henman, 2020), our analysis suggests that institutional memory and historical experience constitute equally important explanatory dimensions.
Existing conceptualisations frequently portray resilience as the capacity of individuals or organisations to adapt to changing environments. In social work, resilience has similarly been associated with adaptation, coping and professional sustainability. However, the present findings suggest that such understandings are incomplete when examining algorithmic governance. Across the four cases, resilience emerged not simply as adaptive capacity but also as the ability to question, negotiate and occasionally resist technological systems that threaten professional discretion and human rights. This interpretation aligns with critical social work scholarship that views professional practice as inherently political rather than purely technical (Abramovitz & Zelnick, 2015), while extending these debates into the field of digital governance.
Our findings also reinforce arguments that digital technologies should not be understood as politically neutral innovations (Latour, 1992; Benjamin, 2019). While the policy discourse frequently frames algorithmic decision-making in terms of efficiency and innovation (Eggers et al., 2017), the comparison illustrates that digital systems simultaneously redistribute power, reshape professional authority and redefine relationships between citizens and the welfare state. In this respect, the study supports recent critiques of algorithmic governance (Eubanks, 2018; Monahan, 2017), while adding a comparative social work perspective that has so far received relatively limited attention.
Recent scholarship has increasingly called for social work to critically examine its own histories of complicity and resistance (Ioakimidis & Wyllie, 2023). Our findings suggest that this historical perspective is equally relevant for understanding digital transformation. Rather than representing an entirely new challenge, algorithmic governance may reproduce longstanding patterns of categorisation, surveillance and exclusion through technologically enhanced mechanisms. Consequently, historical reflexivity becomes an important component of professional resilience.
The comparative analysis demonstrates that AADM is reshaping welfare governance in ways that extend far beyond technological innovation. Across the four national contexts, automated algorithmic decision-making has emerged as a historically and politically embedded mode of governance that reconfigures professional discretion, institutional accountability, and relationships between the state and its citizens.
This study contributes to the emerging literature by demonstrating that resilience is not simply the capacity of social work to adapt to digital transformation; it rather involves the ability of professionals, organisations, and welfare systems to critically engage with, influence, and where necessary resist technological developments that threaten human rights, democratic accountability, and professional ethics. Historical experience, institutional arrangements, and professional cultures all shape how such resilience develops and is sustained.
The findings suggest that the future of social work will depend not only on technological competence but also on the profession’s ability to participate actively in the governance of digital welfare systems. This requires strengthening ethical, legal, and political literacy alongside digital capabilities, while ensuring that algorithmic systems remain transparent, accountable, and subject to democratic oversight.
Ultimately, the central question is not whether social work can adapt to artificial intelligence, but, more importantly, whether increasingly digital welfare systems can remain sufficiently resilient to protect human rights, preserve professional judgement, and maintain democratic accountability. Social work has an essential role to play in shaping this future, ensuring that technological innovation serves social justice rather than simply administrative efficiency.
Goetz Ottmann: conceptualisation, writing – original draft, supervision, methodology, writing – review and editing.
Mara Sanfelici: data curation, investigation, writing – original draft.
Carolyn Noble: data curation, investigation, writing – original draft.
Ines Martinez Herrero: data curation, investigation, writing – original draft.
Caroline Bald: data curation, investigation, writing – original draft.
Eglė Šumskienė: data curation, investigation, writing – review and editing.
Melisa Campana: data curation, investigation, writing – original draft.
Andres Arias Astray: data curation, investigation, writing – original draft.
Renata Nunes: data curation, investigation, writing – original draft.
Abramovitz, M. & Zelnick, J. (2015). Privatization in the human services: Implications for Direct Practice. Clinical Social Work Journal, 43(3), 283–293. https://doi.org/10.1007/s10615-015-0546-1
Agamben, G. (1998). Homo Sacer: Sovereign Power and Bare Life. Stanford University Press.
Bagdonas, A. (2001). Practical and Academic Aspects of Social Work Development in Lithuania. Socialinė Teorija, Empirija, Politika ir Praktika, 1, 35–48. https://doi.org/10.15388/STEPP.2001.0.8495
Bald, C. (2023). Modernisation or Mistake?: Exploring the impact of Covid-19 disruption on School of Health and Social Care course admissions in one university in England and implications for practice readiness. The Journal of Practice Teaching and Learning, 20(1). https://journals.whitingbirch.net/index.php/JPTS/article/view/1966
Ballantyne, N. (2017). husITa celebrates its thirtieth anniversary. https://www.husita.org/husita-celebrates-its-thirtieth-anniversary
Barber, H. (2024, Aug 1). Argentina will use AI to ‘predict future crimes’ but experts worry for citizens’ rights. The Guardian. https://www.theguardian.com/world/article/2024/aug/01/argentina-ai-predicting-future-crimes-citizen-rights
Benjamin, R. (2019). Captivating technology: race, carceral technoscience, and liberatory imagination in everyday life. Duke University Press.
Buck-Morss, S. (2008). Theorizing Today: The Post-Soviet Condition. Log, (11), 23–31. http://www.jstor.org/stable/41765180
Cobo Romero, F., Del Arco Blanco, M. Á., & Ortega López, T. M. (2011). The Stability and Consolidation of the Francoist Regime. The Case of Eastern Andalusia, 1936-1950. Contemporary European History, 20(1), 37–59. http://www.jstor.org/stable/41238342
Edwards, L., Williams, R., & Binns, R. (2021). Legal and regulatory frameworks governing the use of automated decision making and assisted decision making by public sector bodies: Workshop briefing paper. Legal Education Foundation.
Eggers, W. D., Fishman, T., & Kishnani, P. (2017). AI-augmented human services: Using cognitive technologies to transform program delivery. Deloitte Insights. https://www2.deloitte.com/content/dam/insights/us/articles/4152_AI-human-services/4152_AI-human-services.pdf
Eubanks, V. (2018). Automating Inequality: How High-Tech Tools Profile, Policy, and Punish the Poor. New York: St. Martin’s Press.
Fargion, S. (2013). Il metodo del servizio sociale. Carocci.
Folgheraiter, F. (2007). La logica dell’aiuto. Fondamenti per una teoria relazionale del welfare. Erickson.
Gartner, Inc. (2024, April 16). Gartner Announces the Top Government Technology Trends for 2024 [Press release]. https://www.gartner.com/en/newsroom/press-releases/2024-04-16-gartner-announces-the-top-government-technology-trends-for-2024
Gov.UK. (2023a). £٢١ million to roll out artificial intelligence across the NHS. https://www.gov.uk/government/news/21-million-to-roll-out-artificial-intelligence-across-the-nhs
Gov.UK. (2023b). The Bletchley Declaration, policy paper. https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023
Gustafsson, M. S. (2022). Integration of RPS in public services: A tension approach to the case of income support in Sweden. In G. Juell-Skielse, I. Lindgren, & M. Åkesson (Eds.), Service Automation in the Public Sector (pp. 109–127). Springer International Publishing. https://doi.org/10.1007/978-3-030-92644-1_6
Gwadz, M. & Ritchie, A. (2018). Technology Trends: Keep a Wary Eye on Artificial Intelligence. Social Work Today, 22(1), 32. https://www.socialworktoday.com/archive/Winter22p32.shtml#:~:text=Social%20workers%20must%20advocate%20for,enforcement%2C%20and%20health%20and%20wellness
Henman, P. (2020). Governing by algorithms and algorithmic governmentality: Towards machinic judgement. In M. Schuilenburg & R. Peeters (Eds.), The Algorithmic Society: Technology, Power, and Knowledge (pp. 19–34). Routledge. https://doi.org/10.4324/9780429261404-3
Hill, K. (2024). Your face belongs to us: A tale of AI, a secretive startup, and the end of privacy. Penguin Random House.
Ioakimidis, V., & Wyllie, A. (Eds.). (2023). Social work’s histories of complicity and resistances: A tale of two professions. Policy Press.
Juell-Skielse, G., Lindgren, I., & Akesson, M. (2022). Service Automation in the Public Sector: Concepts, Empirical Examples and Challenges. Cham: Springer Nature.
Keddell, E. (2019). Algorithmic Justice in Child Protection: Statistical Fairness, Social Justice and the Implications for Practice. Social Sciences, 8(10), Article 281. https://doi.org/10.3390/socsci8100281
Latour, B. (1992). Where are the missing masses? The sociology of a few mundane artifacts. In W. E. Bijker & J. Law (Eds.), Shaping Technology/Building Society: Studies in Sociotechnical Change (pp. 225–258). Cambridge, MA: MIT Press.
López Peláez, A., Pérez García, R., & Aguilar-Tablada Massó, M. V. (2018). e-Social work: building a new field of specialization in social work? European Journal of Social Work, 21(6), 804–823. https://doi.org/10.1080/13691457.2017.1399256
Mann, M. (2020). Technological politics of automated welfare surveillance: Social (And data) justice through critical qualitative inquiry. Global Perspectives, 1(1), Article 12991. https://doi.org/10.1525/gp.2020.12991
Martinez Herrero, I. (2024). Luces y sombras del trabajo social en la dictadura franquista (España, 1939-1975): Una historia aún por contar. Propuestas Críticas En Trabajo Social-Critical Proposals in Social Work, 4(7), 45–66. https://doi.org/10.5354/2735-6620.2024.72297
Mbembe, A. (2019). Necropolitics. Duke University Press.
Ministero del Lavoro e delle Politiche Sociali (MLPS). (n.d.). Piattaforma per la gestione del Patto per L’inclusione sociale, Ministerio del Lavoro e delle Politiche Sociali. https://pattosocialerdc.lavoro.gov.it/
Monahan, T. (2017). Regulating belonging: Surveillance, inequality, and the cultural production of abjection. Journal of Cultural Economy, 10(2), 191–206. https://doi.org/10.1080/17530350.2016.1273843
OECD. (2025). Government at a Glance 2025. OECD Publishing. https://doi.org/10.1787/0efd0bcd-en
Ottmann, G. (2020). Surveillance, sanctions, and behaviour modification in the name of far-right nationalism: The rise of authoritarian ‘welfare’ in Australia. In C. Noble & G. Ottmann (Eds.), The Challenge of Right-wing Nationalist Populism for Social Work: A Human Rights Approach (pp. 135–150). Routledge. https://doi.org/10.4324/9780429056536-11
Ottmann, G. (2026). Automated algorithmic governance in the human services. In G. Ottmann & C. Noble (Eds.), AI and the Disruption of Welfare: Challenges for Social Work Education and Practice. Routledge.
Palkova, A. (2024). The Baltic states twenty years after the EU’s “Big Bang” enlargement: Political, economic, and social transformations. Romanian Journal of European Affairs, 24(2), 71–96. https://rjea.ier.gov.ro/wp-content/uploads/2024/12/RJEA_vol.24_no.2_Dec-2024-Art-4.pdf
Pavolini, E., & Saraceno, C. (2022). 2021: A year of transition for social security and welfare policies in Italy? Contemporary Italian Politics, 14(2), 260–274. https://doi.org/10.1080/23248823.2022.2064129
Petružytė, D., Gevorgianienė, V., Charenkova, J., Seniutis, M., Šumskienė, E., Žalimienė, L., & Yamaguchi, M. (2023). Integrating technology into social work practice and study: Needs, challenges, and opportunities. Acta Paedagogica Vilnensia, 50, 23–36. https://doi.org/10.15388/ActPaed.2023.50.2
Read, R., & Thelen, T. (2007). Social security and care after socialism. Focaal, 2007(50), 3–18. https://doi.org/10.3167/foc.2007.500102
RTVE. (2025, April 12). El franquismo, nueva “moda indie” para muchos jóvenes que ignoran la realidad de la dictadura. https://www.rtve.es/noticias/20250412/franquismo-nueva-moda-indie-para-muchos-jovenes-ignoran-realidad-dictadura/16538237.shtml
Sanfelici, M. (2024). Fondamenti del servizio sociale anti-oppressivo. Carocci.
Schuilenburg, M. & Peeters, R. (2021). Chapter 1. The algorithmic society: An introduction. In M. Schuilenburg & R. Peeters (Eds.), The Algorithmic Society: Technology, Power, and Knowledge (pp. 1–15). Abingdon and New York: Routledge.
Šumskienė, E., Martínez Herrero, I., & Bald, C. (2026). Response to the special issue ‘Social Work and Social Control’: A Delphi study connecting the coercive past and artificial intelligence futures of social work in Lithuania, Spain and the UK.Critical and Radical Social Work, 1–18. https://doi.org/10.1332/20498608Y2025D000000119
Taylor, A. (2017). Social work and digitalisation: bridging the knowledge gaps. Social Work Education, 36(8), 869–879. https://doi.org/10.1080/02615479.2017.1361924
Valentine, S. (2019). Impoverished Algorithms: Misguided Governments, Flawed Technologies, and Social Control. Fordham Urban Law Journal, 46(2), 364–427.
World Bank. (2023). Il patto per l’inclusione sociale del reddito di cittadinanza: Una valutazione di processo della presa in carico. World Bank, https://www.lavoro.gov.it/sites/default/files/redditodicittadinanza/Documents/Rapporto-Valutazione-di-Processo-Pais-RdC.pdf
Weatherall, K., Henman, P., Bello y Villarino, J.-M., & Matulionyte, R., Sleep, L., Trezise, M., Van Der Arend, J., & Wilcock, S. (2024). Automated decision-making in New South Wales: Mapping and analysis of the use of ADM systems by State and Local governments. ARC Centre of Excellence for Automated Decision-Making and Society. https://doi.org/10.60836/40EB-WE52