Conference Agenda

Platform work is one manifestation of shifts driven by digitalisation. There is a growing body of empirical research on the incidence and characteristics of platform work, as well as concerns raised about the grey area which platform work inhabits in terms of the employment relationship. However Purcell & Brooks (2022:391) note that there has been little accompanying theoretical debate, which they argue ‘has had consequences for the quality, form and potential of debates about platform work to date’.

Some key elements of digital labour platforms, such as algorithmic management practices and digital monitoring, are seeping through to conventional work settings across different sectors and countries, and likely to grow in the future (Urzi-Brancati et al. 2020; Woods 2021; Baiocco et al. 2022, Fernández Macías 2023, Rani et al. 2024). In addition to concerns about the potential intrusiveness of digital tools and how constant surveillance may affect workers’ wellbeing, the breaking down of jobs into smaller tasks required by algorithmic management may lead to an increasing fragmentation and commodification of labour (Franke and Pulignano 2021; Ball 2021). Thus, a repositioning of the traditional frames of the labour process is required. This ‘platformisation’ of work can be deepened and accelerated by recent developments in artificial intelligence which will fundamentally affect work processes. These implications may well require changes in the current labour market institutions and regulations.

To measure, better understand and characterise the transformation of both content and work processes as a result of the introduction of algorithmic management, digital monitoring and AI, we introduce the AIM-WORK survey which is representative of the working age population in all EU countries. Empirical data from AIM-WORK is used to measure the prevalence and characteristics of digital monitoring, algorithmic management and platform work in the EU, and investigate how they are being used to organise, coordinate and control platform workers and regular workers alike. Preliminary empirical analysis suggests that platformisation of work is a real phenomenon affecting the labour process in a relevant proportion of workers in the European context. Additionally the analysis delves deeper into the implications of AI tools on the world of work. Our analysis underpins a better theoretical understanding on the impact of digital technology on the labour process.

The aim of this joint research project by ILO and NASK is to develop a system of more precise assessment of potential effects of generative AI (GenAI) on labour markets. We build on the methods developed by the ILO (Gmyrek, Berg, and Bescond 2023) and further refined in cooperation with the World Bank (Gmyrek, Winkler, and Garganta 2024), fitting into the rapidly emerging literature on modelling of potential impacts of AI on tasks and occupations (Acemoglu, 2024; Acemoglu et al., 2024; Acemoglu and Restrepo, 2022; Svanberg et al., 2024, Comunale and Manera, 2024 for literature review).

By using the official national 6-digit classification of occupations in Poland, we expand tenfold the number of tasks compared to the studies conducted solely on the ISCO-08 structure (Gmyrek, Berg, and Bescond 2023). We capture the opinion of 1650 people currently in employment in each ISCO-08 1-digit group to rank automation potential of a representative sample of tasks and combine this with a detailed review by a mix of national and international experts. We then train an AI model to reflect human judgement and re-generate scores for the tasks in ISCO-08, previously provided by Gmyrek et al. (2023). This leads to an adjustment of the 2023 ILO index of GenAI exposure, and an update of global, regional and income-based employment estimates. We also run predictions based on some 30,000 6-digit tasks in Poland’s official occupational classification system and provide detailed employment estimates and characteristics of affected demographics in the national context, based on ILO’s micro data.

Our study responds to the need for modelling of GenAI’s potential impact, given the growing societal angst and the recent reports about low levels of preparedness (Bick, Blandin, and Deming 2024; Maison & Partners and ThinkTank 2024). While several indicators of occupational exposure to broader AI technologies are available (Nurski and Vansteenkiste 2024), very few tools focus on GenAI (Nurski and Ruer 2024). Our new, more precise exposure index enables a closer alignment of the academic work on digital economy with the significant interest in the GenAI technology in the public debate. The methodological blueprint provided by our research uses open-source code and can be flexibly expanded to other types of AI or focused more narrowly on specific subsets of digital technologies and sectors. This paper is first in the series of publications, which will also include qualitative elements and a nationally representative survey on GenAI workplace adoption.

In recent years, the debate on the use of artificial intelligence (AI) has gained increasing momentum. This also applies to public administration and, in particular, to public employment services. Placement officers perform a central function on the street level of the labor market. They assess job seekers’ qualifications to identify a profile, develop search strategies and make job offers. They provide advice and guidance, and help job seekers to find further training that extends their competences meaningfully.

Algorithmic decision support systems (ADSS) offer opportunities to help placement officers to fulfill their consulting and mediation tasks. The productive use of algorithmic decision support systems (ADSS) depends heavily on user acceptance of the systems. We are therefore interested in what attitudes employees in public employment agencies generally have towards the use of ADSSs and whether the extent of acceptance differs between specific tasks carried out by placement officers.

The design of an ADSS requires normative choices on its (1) safety, (2) accountability, (3) transparency, and (4) efficiency. While some studies in this field of research are based on qualitative research designs, we worked with a sample of potential future users of ADSS in the public employment agencies and carried out a standardized survey. In particular we use a conjoint experiment to explore the role of different design features for the acceptance of a hypothetical ADSS that uses artificial intelligence to identify individual needs for further training and vary concerning the four normative features.

For our survey we addressed 5,000 randomly drawn placement officers from a total of 150 public employment agencies in Germany. All of them got an invitation e-mail asking them to participate in the survey. The e-mail also contained a short description of the topic, a reference to the voluntary nature of the study, and a link to the relevant information concerning data protection regulations. After data cleansing, we used a net sample of 1,415 individuals for further analyses.

The empirical results indicate that placement officers’ balance privacy and safety standards while maintaining professional accountability. They appreciate ADSS’s support but firmly reject the mandatory use of such advice.

In many Western industrial societies, work has historically represented more than a productive economic activity; it has served as a key factor in personal identity and social cohesion. Digital automation emerges as a possible element of disruption to this foundation, raising critical questions about the role of work in a future where productivity can be increasingly decoupled from human effort in the traditional sense, challenging long-held notions of skills, merit, and value in the workforce. This study examines the need to reassess labor institutions and the evolving role of work, questioning whether the European Industry 5.0 framework – centered on human-centricity, resilience, and sustainability – offers a suitable approach. Examining transformations in the labor market through the lens of AI ethics and drawing on critical questions and debates from political philosophy and social theory enables a broader assessment of the societal implications of automation. In addition to theoretical analysis, the research incorporates an evaluative study of Industry 5.0 as an alternative framework for addressing these challenges. To evaluate the economic feasibility of the Industry 5.0 framework, the study analyzes a selection of case studies presented in the context of the Business Case 5.0 for Sustainable Competitiveness, within the Community of Practice 5.0, promoted by the DG RTD of the European Commission. Contextually, this research situates itself within ongoing debates on both “meaningful work” and “post-work” perspectives. On one hand, it engages with arguments that wage labor serves as an essential source of meaning, re-examining historical frameworks that link work to social cohesion and reassessing the role of labor in fostering human and social connections. On the other hand, in line with discussions on the future of European competitiveness, the study evaluates the viability of a human-centric policy framework for industrial transformation. For automation to contribute positively to human well-being, a renewed societal agreement should emphasize shared responsibility among governments, industries, and citizens and, within this framework, policies that promote equitable access to education, reskilling opportunities, and social protections should play a key role in shaping its development. Moreover, a more expansive understanding of productivity – one that reflects psychological, social, and environmental well-being – could serve as a foundation for this transition. By engaging with its principles, this study evaluates the potential of Industry 5.0 to serve as a normative counterweight aligning technological progress with holistic human flourishing, while noting the gap between conceptual aspirations and implementation.