Introduction to the Special Issue on Activation and the Digital Welfare State

Employment services have been subject to repeated reform for 30 years, particularly in Australia. The country has been at the forefront of welfare-to-work reform trends internationally, pioneering several of the public management reform agendas that have reshaped the governance of activation since the 1990s (van Berkel and Borghi 2008). These include the contracting-out of case management of the long-term unemployed in the early 1990s, the move to full marketisation in the late 1990s, followed by full privatisation in the early 2000s and an ever-increasing emphasis on results-based payment models (O'Sullivan et al. 2021). Broadly reflecting a governance approach anchored in the ‘New Public Management of activation’ (Weishaupt 2024), these reform dynamics—dubbed a project of ‘double activation’ (Considine et al. 2015)—have unfolded hand-in-glove with substantial changes in social policy designed to ‘activate’ both claimants and providers. On the one hand, jobseekers are ‘activated’ into work, principally through ‘demanding’ (Raffass 2017) behavioural instruments like onerous job-search requirements, Work-for-the-Dole and other mutual obligations. On the other hand, successive waves of administrative reform have also involved subjecting providers to many of the same behavioural policy instruments (targets, performance monitoring, financial incentives) as claimants, so as to motivate them to activate jobseekers ‘in the right way’ (Andersen and Larsen 2024: 8). An important contribution of the analytic lens of double activation is that it directs enquiry towards the political effects of seemingly technical reforms. This is through examining the distributional consequences of administrative reforms for claimants' access to social rights and surfacing the ways in which adjustments to how policies are delivered reconfigure what policies are enacted (Brodkin 2013). This is an underlying theme of this special issue, which focuses on the latest developments in the governance of welfare-to-work amidst a period of unprecedented ‘digitalisation’ in welfare administration (Alston 2019; Busemeyer et al. 2022; Henman 2022, 2025). The past decade has seen a ‘“digital turn” in the governance of welfare’ (Papadopoulos and O'Keeffe 2023). Encounters with the welfare state now increasingly unfold through ‘digital interfaces’ (Meers et al. 2024) as data analytics and machine learning technologies are increasingly deployed to profile claimants and target the delivery of activation in more granular ways (Carney 2021; Henman 2022; Pors and Schou 2021). Ongoing reforms in Australia exemplify these dynamics. In July 2022, Australia introduced a new Workforce Australia model whereby a substantial proportion of jobseekers now self-manage their mutual obligations via an online portal with minimal human contact (Workforce Australia Online). Traditional employment support services provided by case managers, or employment consultants, continue to be delivered by contracted providers delivering what is known as Workforce Australia Services. However, this in-person support has been refocused towards participants ‘who have greater barriers to employment and require more intensive support’ (Nous Group 2024: 22), such as those with mental health issues, drug or alcohol addiction, and risk of homelessness (among other factors). The underlying logic is that ‘efficiencies from the reduction in face-to-face servicing arrangements’ can be reinvested ‘to provide a more intensive, targeted and tailored service for those who need extra help’ (Auditor-General 2023: 19–20). As of 31 August 2024, just under 175,000 of the 638,778 participants in Workforce Australia were self-servicing online (DEWR 2024). Whether jobseekers self-service or receive in-person support is largely determined by how they respond to a series of assessment questions to differentiate the ‘job ready’ and digitally literate from higher-needs claimants (Papadopoulos 2025). Underpinning both Workforce Australia Online and Workforce Australia Services is a new points-based activation system (PBAS) purporting to offer greater flexibility and choice over how claimants meet their mutual obligations. Instead of the standardised 20 job searches per month used under previous systems, participants can now accumulate points for a multiplicity of activities, including volunteer work, online training, gaining industry certifications, and attending job interviews. It is expected that the introduction of the PBAS and refocusing of in-person services towards those with more complex needs will see a shift in emphasis away from a ‘work-first’ activation model. Indeed, Workforce Australia has been explicitly framed by the government as a ‘departure from the “work-first” philosophy that underpinned jobactive’ (DEWR 2023). However, this is much debated. Online self-servicing has been likened by critics to a regime of ‘digital dole parole’ (Casey 2022). The chair of the recent Select Committee inquiry on employment services described the new jobseeker compliance framework as ‘Boston Consulting meets Squid Game in a dystopian reality TV show that we make citizens play every day’ (Hill 2022). A key concern is the automated nature of payment suspensions if self-servicing participants fail to upload documentation in time. While Australia has gone further than most in automating services, it is far from alone. Other countries at the vanguard of this digital turn include Belgium and Norway, where agencies have been using digital technologies for several years. In Belgium, the Flemish employment service (VDAB) has developed Artificial Intelligence (AI)-based jobseeker profiling and job-matching tools to triage jobseekers for more intensive, face-to-face support and encourage participants to apply for local vacancies they may not have considered but which match their competencies according to an algorithmic reading of their résumé (Desiere and Struyven 2021). In Norway, the public employment service (NAV) likewise uses a digital profiling tool to support advisors determining what intensity of support to provide jobseekers. Before accessing services, jobseekers must first register online and complete a questionnaire that is algorithmically processed to produce ‘an expected level of need’ (Løberg 2023: 21). In a similar way to the Job Seeker Snapshot, advisors make initial determinations—based on this questionnaire and without meeting with clients—about whether participants can manage by themselves or require long-term support. NAV also uses online channels to support interactions between caseworkers and clients via ‘digitally mediated service encounters’ (Breit et al. 2021: 835). NAV's Modia system integrates a chat platform with a digital activity plan to enable advisors to remotely communicate tasks to jobseekers while jobseekers can similarly log their completion of activities from a distance. The Online Journal and Online Account for Universal Credit claimants in the UK perform a similar role (Bennett et al. 2024). However, unlike Workforce Australia Online, human advisors continue to drive how citizens are being activated in the Norwegian and UK systems, with digital interfaces being deployed not to replace caseworkers but as ‘alternatives to physical interactions between caseworkers and clients by means of digital layers’ (Scarano and Colfer 2022: 101). This underscores the importance of differentiating between the multiplicity of forms that digitalisation in welfare administration takes. Proclamations of the advent of a new digital era of welfare governance (Alston 2019; Yeung 2023) must be sensitive to distinct ‘varieties of digitalisation’ (Considine et al. 2022: 521) that are unfolding and how they affect claimants and workers in varied ways (see also Carney 2020). For instance, Scarano and Colfer (2022: 101) differentiate digital interfaces that facilitate ‘remotisation’ from the deployment of datafication and machine learning for purposes of ‘automisation’. The latter involves deploying data integration and predictive analytics to generate new forms of knowledge about participants that can be used to rationally target interventions. The former are orientated towards enabling caseworkers and clients to interact mostly as before, but through virtual meetings and chat functionalities. Ball et al. (2023) expand on this typology to differentiate between three models of digitalisation in welfare-to-work: virtual engagement, or remote activation, as in the examples of Norwegian and UK employment services; transactional automation, or self-activation, as typified by Workforce Australia Online; and digital triaging, or targeted activation, which is most advanced in Belgium. Notwithstanding its varied modes, the digital turn in the delivery of activation is substantially reshaping how welfare-to-work policies are being enacted on the ground and experienced by citizens at the ‘street-level’. This is through its effects on citizens' experiences of administrative burden (Ball et al. 2025; Bennett et al. 2024) and its implications for the ‘classical model’ (Pedersen and Wilkinson 2018: 195) of service provision by frontline workers. Administrative burden is a framework for understanding the variety of ways in which citizens' access to rights can be undermined by cumbersome and challenging bureaucratic processes and the costs they might incur when trying to access rights and services. In the literature, particular attention has been paid to three categories of costs: learning costs, such as the time and effort that goes into learning about programmes or regulations; compliance costs in the form of money, time, and effort spent on providing documentation to prove eligibility and responding to the ‘discretionary demands made by administrators’ (Herd et al. 2023: 4); and psychological costs such as experiences of stigma, stress, and autonomy loss resulting from bureaucratic encounters (Herd and Moynihan 2018). For technology-savvy participants with smart devices and internet bandwidth, digitalisation reforms may ameliorate some of the psychological and compliance costs involved in welfare-to-work participation. The ability to log evidence of the completion of job-search and other activity requirements remotely and 24/7 access potentially saves jobseekers time from having to travel to employment services offices and wait in line for appointments (a substantial compliance cost). Activation becomes less demanding when claimants can schedule their completion (and verification) of behavioural requirements around other life demands, rather than being dictated by the rhythms of office opening hours and caseworkers' diaries. Likewise, being able to lodge documentation online may be experienced as less stigmatising than having to publicly present to dole offices or job centres to prove eligibility for payments. Furthermore, in countries with dispersed populations and limited physical infrastructure, the delivery of services through digital platforms may significantly widen the reach of social security payments and employability services, assuming the availability of stable internet. However, critics point to several ways in which digitalisation may intensify experiences of administrative burden. One way is through systems of self-activation being directed towards ‘strict[er] delivery of the conditions under which welfare is accessed’ (Morris et al. 2020: 27). For instance, the shift towards digital job plans (activity agreements) and algorithmic processing of payment suspensions may expose claimants to more onerous behavioural demands and a higher risk of incurring sanctions. From this perspective, one concern regarding automated implementation structures is the risk of digitalisation being deployed to embed ‘compliance … as the dominant operating objectives of activation’ (Papadopoulos and O'Keeffe 2023: 14) through more consistent application of stringent conditionality requirements. Moreover, systems of self-activation potentially ratchet up learning costs for participants in terms of navigating the administrative hoops for claiming payments and accessing information about supports they are entitled to. As Bennett, Currie, and Podoletz highlight in the context of the UK's digital-by-default system for claiming payments, it now falls on claimants to about the 14) and to trying to administrative and payment when systems produce The enabling of activation, such as access to and support for may also through processes of not they are but in a system clients are jobseekers must what they what is if their needs are not being and what they (Ball et al. The that when systems to digital is that it is by the of physical opening hours or office clients can questions and the is as the algorithmic processing of can also make the of administrative less (van and Carney 2024). This is to the of with caseworkers of and state and of to claimants about the that affect Carney terms this of administrative and its importance is by the automated to claimants under the many of understanding of how they a 2022: and were by the they trying to contact This important questions about the between digitalisation reforms and the to which the implementation of activation will continue to be by For a time, frontline in welfare offices have been considered of advanced and This from their role in social policies into activation (van Berkel 2020). including this as an of policy implementation that is to the of services that are to the of citizens it also frontline workers a role in access to social rights and experiences of activation subject to the of this claimants the same time, and may one but many welfare on which office they and which they the of frontline over policy implementation has been a concern of waves of administrative as by the project of double activation 2021). This project new in the digital welfare where the of automation, as as and algorithmic forms of are framed as the between policy and policy The of algorithmic profiling for is as case less to and in (Scarano and Colfer 2022: The is a more of through greater in service and the of case management towards those with minimal support However, this is to that the processing of about service can be ‘to an and (Pedersen and Wilkinson 2018: It that bureaucratic encounters are more than transactional application of (see Carney rather than encounters by that and (O'Sullivan and 2018: are of administrative in or through of (Considine et al. 2022: The to which algorithmic profiling tools can enable services to be targeted more on several is their predictive in those most from the from the and 2023: This is with several the of profiling tools to participants and to (see 2018). For instance, a of profiling tool that those Belgium were as as their to be as being at risk of long-term (and subject to more intensive (Desiere and Struyven 2021). about the predictive of profiling tools are not as algorithmic assessment tools have been used to target services for and 2023). However, the of these instruments to participants towards self-activation the consequences of as jobseekers may have with advisors who can (Ball et al. 2025). In most algorithmic profiling tools have been deployed as rather than to replace caseworkers' a further when their becomes the to which advisors to or on their is so that advisors are of the of algorithmic profiling tools and them (Løberg 2023). that when advisors are about the predictive of most are to their for of the and 2022). this is the having a is not the administrative and that many it It may be a but it will not be a to citizens when machine This special both the and by the digital turn in employment services, both for the workers involved in policy implementation and claimants' experiences of It from a on The of at the of in with support from the of of Australia. were first at that with through an for While several have been on welfare digitalisation more and this is the first to how processes of digitalisation are reshaping the of how processes of digitalisation are reshaping activation and citizens' experiences of welfare-to-work has to been A of this is to and this but of to the ways in which digitalisation reforms are reshaping the delivery of welfare-to-work and the between demanding and enabling of The with an early of the of digital technologies by providers in the UK and Australia. et al. with to how providers in both countries were using digital tools in the to the of new In both digital tools to be used for and remote servicing rather than as means of This by the providers that they to some of service delivery in of for Indeed, a the to which particularly in using digital tools for their compliance with requirements and for the performance of This underscores the double activation logic to which providers are with and how digital tools of this governance online not instruments for compliance with mutual obligations but also a means of the reach of of performance and that the of digitalisation at a more advanced in Australia than the the new Workforce Australia However, concern many providers that the digital turn and that plans for expose participants to new of digital if were In their et al. the policy towards jobseekers and of claimants' that are to the of employment services. this in caseworkers' and of activation, which they are to in digital how these are to the deployment of they on data from as of an project to a digital job-search support tool for a employment The were countries with of and unemployed were with a of the tool and to their on purposes for which it be involved deploying the tool to provide information to jobseekers about the state of the and their of For instance, using the tool to provide jobseekers with an algorithmic of how it them to the tool to present claimants with information about of in their local or about the and security of the that they were being to apply This of information is to jobseekers by employment services providers or administration the and that caseworkers and were about the that deploying digital job-search tools in such ways have on claimants' were as in need of from the reality of their job they This the of activation work, which the through a of the between and the of points towards a of the of digital where the of delivery becomes by the of and ‘work-first’ logic of The three subject employment services model and the technologies to intensive In the deployment of algorithmic profiling tools in the of activation through a of the Job Seeker on of inquiry and government the and of the and purposes of the over a tool for claimants' of intensive to a tool for determining payments in a welfare-to-work into a digital tool for face-to-face services. attention to how the logic of profiling on experiences of into employment In so the repeated in how the of profiling is as a for long-term or a tool for support policy have between these but of risk not of support The predictive of the in claimants' of long-term unemployed is of its as a this latter being increasingly important to how it is being used to digitally triage participants in Workforce Australia. a granular case of the technologies of self-activation Workforce Australia Online, employment services as an of welfare reform by the of social security and of of service and inquiry how the devices of the digital job plan and online of activity requirements have a new regime of This is as online job plans participants into a and standardised implementation of points-based activation, as by the nature of the activity participants and the of online that is digitally into that a ‘work-first’ logic has been into Workforce Australia Online by the of job as the activity and the of the system to algorithmically about the application of payment suspensions and over and conditionality has from the level to the the for participants to or payment suspensions and Furthermore, attention to the that can the of the that automated more of that systems of self-activation may to new of administrative The theme of administrative is through and which how frontline workers about their can to the implementation of administrative that and on their and Ball their through a case of the frontline of conditionality requirements under the which has been in for its effects on and mental health et al. the but processes of and that play in frontline of welfare conditionality in so to social policy on administrative with public policy on administrative burden. The the Australia to include and that have digital technologies to new welfare-to-work amidst and Scarano Employment which to digitalisation in three to access to to participants to activation and to the and job competencies of delivery However, and Scarano a between the policy and the reality of that the from digital For instance, they that the of online platforms for delivering services such as appointments is ‘remotisation’ not administrative for participants of requirements to key compliance documentation in such as to must be online but in of The is a of bureaucratic rather than administrative and a refocusing of encounters towards more also questions about the of the used to participants to activation participants complete an online questionnaire that is used to their risk of long-term This is by the administration of a standardised assessment caseworkers have minimal to the assessment over of Employment participants to have been to the ready’ support being from Furthermore, are with the that may be while profiling instruments may also be important Scarano and highlight the of regarding the profiling and of et al. a to how digitalisation access to on an from the It is one of the of welfare digitalisation the of welfare and an of the for digitalisation to facilitate more enabling forms of In the this through digitalisation to to dispersed that not have access to access to payments through an online rather than physical made it and less to payments and to in (a for These services were in turn provided through online with the government digitalisation with of marketisation through a where be from several of contracted providers the digital as as delivered the online This approach choice over a choice that further by the online of delivery access to Notwithstanding these important and choice and highlight an important of the implementation that digitalisation access towards participants from more while those with issues, limited and A further concern the by the platform which and in terms of how were to and that the implementation of an important of how the digital turn in welfare administration can to widen access to payments and services in countries limited financial and administrative et al. 2024: The and of the digital turn in the governance of activation are It is that in social policy and public administration has to these developments through and of the implications for frontline and the administrative citizens when accessing as by the to this special issue, ‘digitalisation’ is not one but has many in the context of welfare-to-work reforms. of these activation in more enabling while ‘work-first’ policy and demanding activation this it is not the technical platforms as such but the political in which they are that are most for how processes of digitalisation are reshaping the of activation and experiences of and and and for this by from an in with the Employment Services and Group The on The of which the of this special issue, a from the of in Australia. The like to the and for to this special and for their on an The of