Abstract
The accelerating digital transformation, and particularly the diffusion of generative AI technologies, poses profound challenges for workplace co-determination. Works councils increasingly face the need to assess and negotiate technological innovations under conditions of epistemic asymmetry, time pressure, and limited experience in (socio)-technical reflection. This paper introduces the Implication Canvas, an adapted version of the Implication Fan developed at the Berlin Ethics Lab, as a participatory tool to strengthen deliberative capacity in co-determination contexts with a focus on socio-ethical reflection. Drawing on an explorative qualitative research design, including participatory observation and post-workshop surveys, the study examines iterative adaptations of the canvas in three workshops with a cooperating works council. Findings indicate that the adapted canvas facilitates structured reflection on preconditions, consequences, and solution pathways of AI-driven workplace innovations while bridging communicative and epistemic gaps between management, technical experts, and employee representatives. The study argues that effective and responsible co-determination should be understood as an integrative competence combining technical orientation, ethical reflexivity, and procedural agency. Participatory reflection tools such as the Implication Canvas can support the practical enactment of this competence and contribute to more inclusive and responsible workplace AI governance.
Similar content being viewed by others
1 Introduction
In today’s technology-driven world, works councils increasingly face the task of evaluating the ethical, legal, and socio-organizational implications of artificial intelligence (AI)-supported digitalization initiatives brought forth by company leadership. Works councils are legally institutionalized bodies of employee representation at the establishment level in Germany. They are elected by the workforce and participate in matters concerning working conditions, personnel issues, and organizational change. While analogous institutions exist in some European countries, the German case is particularly instructive because co-determination rights are comparatively well formalized and extend into processes of workplace digitalization. This makes the German context a useful site for studying how employee representatives can respond to the rapid diffusion of AI in organizational practice. In 2021, the Works Council Modernization Act (Betriebsrätemodernisierungsgesetz, 2021) extended the works councils’ rights to co-determine matters specific to digitalization and the introduction of AI in the workplace. Most significantly, the act treated the introduction of AI as a matter that automatically warrants consulting external expertise—a condition that previously and in other matters required the employer’s agreement. Hence, to a certain degree, the German law already recognizes the complexities of AI-based transitions in the workplace and the fact that works councils may not be in a position to effectively co-determine in these matters—as a rule, not as an exception. In a joint projectFootnote 1 with the works council of a municipal utility provider, the evolving challenges of workplace co-determination in the context of digital transformation were addressed, with a particular focus on the adoption of generative AI. The project was carried out in Germany, where there are particularly strong legal regulations aimed at securing employee co-determination in all matters concerning personnel, social aspects, and corporate change processes. Against this background, the present study addresses the following research problem: how can works councils be supported in developing structured, practice-oriented, and socio-ethically informed judgments about AI-related workplace transformations under conditions of uncertainty and asymmetrical expertise?
To address this problem and in order to empower council members to undertake such independent assessments, a structured evaluation instrument was developed. This instrument, which has been termed the Implication Canvas, was developed through an iterative process with works council representatives, following an exploratory qualitative research design. The objective of this tool is to facilitate informed, efficient, and autonomous decision-making. Furthermore, the aim of the tool development is to establish a structured yet open framework for collective reflection that enables the systematic integration of diverse perspectives. This approach is motivated by the observation that a somewhat unstructured culture of discussion sometimes prevails in co-determination bodies, due to varying levels of expertise and areas of responsibility, as well as to member turnover, since members are typically re-elected every four years. By making potential consequences explicit, surfacing underlying assumptions, and supporting the formulation of action strategies, the tool seeks to reduce epistemic asymmetries and to strengthen collective judgment in co-determination processes. Rather than producing definitive evaluations, the Implication Canvas is designed to open a deliberative space in which reflective and forward-looking co-determination can be practically enacted and continuously developed as a shared organizational capability. The instrument is embedded in a modular workshop framework designed to balance openness and procedural guidance with co-creation approaches, thereby enabling participants to critically examine technological innovations, deliberate collectively, and develop context-sensitive response strategies. We provide insight into the development of this tool as a flexible workshop concept designed to address diverse socio-ethical challenges. The conceptual foundations draw on the educational framework of “Challenge-Based Learning” (CBL) (Nichols & Cator, 2008) which emphasizes real-world cases, learners’ agency in identifying problems, and interdisciplinary environments. CBL is also often used in global and technical contexts (cf. Malmqvist et al., 2015) and originated in an industrial setting. This approach has already been translated to and implemented in higher education and refined through multiple iterations (cf., Herzog et al., 2022). This stands in contrast to checklist-based approaches, in which relevant normative categories are treated as fixed. CBL is particularly suitable in this context because it conceptualizes learning as situated in authentic problem contexts and driven by participants’ active sense-making, which closely mirrors the case-based, experience-oriented, and collectively reflective character of co-determination practices. This tradition ensures a strong grounding in real industrial challenges and adds a systematic emphasis on participants’ autonomous construction of models and evaluative perspectives, which can be crucial for developing transferable ethical reasoning capacities within co-determination processes. On this basis, the structure of the “Implication Fan” (Fischer & Mehnert, 2021) is recovered and adapted to the deliberative setting of works councils. In order to elucidate the relationship between the three central elements of our approach more tangibly, the Implication Fan is the original reflection tool developed at the Berlin Ethics Lab to map the prerequisites, consequences, ethical issues, and interdependencies of socio-technical innovations across different time horizons. The Implication Canvas is the practice-oriented adaptation developed in this project for use in co-determination settings. CBL provides the pedagogical framework that guided the modification of the original tool: moving away from an academic-led reflection format and toward a participatory, action-oriented, and case-based process in which participants collaboratively develop an understanding of the problem and devise response strategies.
We report empirical insights from applying the tool in ongoing co-determination processes, in collaboration with the works council of a municipal utilities provider. The council was selected due to the company’s variety of business areas, which include energy supply, telecommunications infrastructure, and public transportation. This diversity mattered analytically because it exposed the workshops to a broader range of occupational settings, digitalization trajectories, and potential AI application scenarios than a more specialized company would have allowed. In addition, the organization is currently implementing multiple AI-based systems, particularly generative AI tools, whose broad accessibility via natural language interfaces expands both the quantitative scope and qualitative nature of potential applications. Accordingly, these developments will require works councils’ members and employees to form well-grounded positions under conditions of time pressure, resource scarcity, and increased expectations for efficiency and service expansion—amid an ongoing shortage of skilled labor (cf., e.g., Butollo et al., 2024; Koch & Lodefalk, 2025; Krzywdzinski, 2024). As Gerbracht et al. point out, the most effective way to strengthen co-determination practice would be to use a tool that addresses the diversity of works councils appropriately (2024, p. 18 f.). As we will substantiate in the sequel, we have therefore designed our concept in such a way that it can be adapted to the specific needs of works councils in a wide variety of practical projects and can be independently customized and further developed. In doing so, the approach is intended to contribute to the development of reflexive digital competencies, that emerge from both formal knowledge and everyday life experience (Stubbe, 2017, p. 48). This, we suggest, can generate multiplier effects within organizations as participants share and disseminate these practices, enabling broader institutional uptake. The approach is intended to be transferable to other co-determination bodies.
In this article, we seek to contribute to three interrelated strands of research. First, we contribute to socio-ethical reflection tools (cf. Lewis et al., 2017; ODISSEI, n.d.; Portegies et al., 2025) and their role in supporting fair and equitable co-determination practices during the digital transformation. Second, we contribute to participatory design (e.g. Dindler et al., 2026; Gerbracht et al., 2024; Ruess et al., 2024) by adopting and translating a CBL approach from education (e.g. Bombaerts, 2021; Doulougeri et al., 2024; Gallagher & Savage, 2023; Herzog et al., 2022) for use in co-determination settings. Third, we contribute to qualitative accounts of constructive co-determination practices in workplace AI governance (e.g. Haipeter et al., 2024; Krzywdzinski et al., 2023; Milanez, 2025), by conceptualizing effective and responsible co-determination as an integrative competence involving technical orientation, ethical reflexivity, and procedural agency. In doing so, the article positions the Implication Canvas as a participatory instrument that operationalizes structured socio-ethical reflection for works councils and thus extends existing CBL-inspired tool development into the field of institutional co-determination.
The remainder of the paper is structured as follows: Sect. 2 outlines the challenges of workplace co-determination under conditions of digital transformation. Section 3 describes the theoretical background. Section 4 details the research design, introduces the original Implication Fan as the epistemic foundation of our approach and describes the workshop settings. Subsequently, the empirical results concerning the adaptations of the workshop and canvas are elucidated in Sect. 5. In Sect. 6, we analyze three of the most pressing challenges identified, considering our results from the development process. Section 7 discusses the results and concludes with implications for practice and avenues for future research.
2 Workplace Co-Determination in the Context of Digital Transformation
The digital transformation poses fundamental challenges to workplace co-determination. Technological innovations, particularly the deployment of generative AI, are reshaping work processes, skill requirements, and organizational structures at a high speed (Bader & Kaiser, 2020, pp. 3–5; Grasy et al., 2024, p. 5; Niewerth et al., 2022, pp. 19, 24; Widuckel, 2020, p. 31). The rapid pace of innovation in digital technologies, especially in generative AI, amplifies this dynamic. New tools and application scenarios emerge in short cycles, often accompanied by strategic expectations regarding efficiency gains and organizational restructuring (Niewerth et al., 2022, p. 25). This transformation creates a field of tension between corporate strategy formation, technological design options, and the safeguarding of employee interests. Accordingly, employees’ interests should be protected so that they are not at the mercy of opaqueFootnote 2, highly complex AI applications that promise to be saviors. The risks of this transformation could include loss of skills (Bainbridge, 1983; Krook, 2025, p. 6011 f.; Véliz, 2024), worker replacement (Frey & Osborne, 2017; Moore, 2019), broader redistribution of tasks and areas of work (Guliyev, 2023; Milanez, 2023, p. 15; Wang & Lu, 2025), and in some areas, excessive demands following the elimination of routine tasks. A recent survey by the ifo Institute (Ruffert, 2026) of “roughly 3,000 businesses” (Aktien Check, 2026) in Germany shows that the adoption of AI is reshaping how firms think about skills and staffing. Nearly one in five companies currently using AI report seeing potential to replace university graduates with less-qualified workers operating AI tools, particularly in the retail and service sectors. While the majority of firms still consider such substitution difficult or impossible, the survey results demonstrate that AI can facilitate deskilling by transforming complex tasks into standardized, AI-supported routines and replacing workers based on their qualification profiles (Ruffert, 2026). In the German context, where co-determination rights are comparatively strong, these developments underscore the need for works councils to anticipate not only direct job losses but also more subtle shifts in how skills and responsibilities are distributed across the workforce. As discussed in Sects. 6 and 7, similar tensions between relief, deskilling, and changing responsibility structures emerged in the Copilot cases examined with the cooperating works council. In this context, it could be argued that works councils should assume a pivotal role as mediators between management, technical experts, and the workforce, acting on behalf of all employees and thus serving as a protective entity for them. However, their capacity to act is structurally constrained by a set of interrelated conditions that shape contemporary co-determination.
A first central challenge (a.) concerns epistemic differences between management and works councils. Strategic decisions regarding digitalization are generally developed within specialized technology and leadership units, while works councils primarily draw on legal expertise and situated knowledge of working conditions.Footnote 3 They rarely have equivalent access to technical expertise (Gerbracht et al., 2024, p. 18), however, they are allowed by law to consult external experts. In some cases, they are unable to call on this expertise sufficiently, presumably due to scheduling commitments, costly consulting fees, or limited access to comprehensively trained experts who could provide more extensive technology impact assessments. As a consequence, technological transformation processes risk producing asymmetric relations of definitional authority, particularly when technical knowledge becomes the principal source of legitimacy for organizational decision-making (Kellogg et al., 2020, p. 367 f.). This can position works councils reactively and narrow their scope for intervention when technology-driven decisions appear predetermined or without viable alternatives. This asymmetry is reinforced by a second challenge (b.): decision-making under time pressure and uncertainty. Works council members typically perform their representative tasks alongside their regular professional duties (Gerbracht et al., 2024, p. 18). Time and cognitive resources for sustained engagement with complex technological developments are therefore limited, while innovation cycles continue to accelerate. Co-determination thus increasingly takes place under conditions of incomplete information and temporal constraint. Beyond these structural constraints, co-determination is increasingly confronted with a third challenge (c.): the assessment of socio-technical implications of AI systems. The task is no longer limited to the evaluation of discrete technical functionalities but requires anticipation of organizational, social, and normative effects, including changes in work processes, qualification structures, and control relations as a kind of ‘future skill’. At the same time, digital transformation is embedded in organizational future narratives (Grasy et al., 2024, p. 40) that promise efficiency gains, service expansion, and strategic modernization. In this context, the adoption of the latest technologies are often presented as an indispensable obligation, accompanied by specific, predefined implementation plans that may not necessarily be in the best interests of the workforce. These narratives constitute a fourth challenge (d.), as they prestructure decision-making processes and render alternative technological pathways less visible, particularly when financial investments, vendor commitments, or managerial expectations have been established. Once technological innovation itself is no longer the subject of intensive scrutiny, path dependencies deprive those involved in workplace co-determination of the opportunity to put forward alternative proposals. Finally, co-determination is confronted with a fifth challenge (e.): the protection of collective labor interests under transformation conditions. Risks such as deskilling (Ferdman, 2025), task redistribution, displacement, and work intensification often unfold indirectly and over time, making them difficult to contest at the point of technology adoption. Rendering these effects visible and negotiable thus becomes a core dimension of contemporary co-determination. Against this backdrop, the development of reflexive and process-oriented forms of judgment becomes a key requirement. Works councils must be able not only to evaluate specific technologies, but also to reconstruct their organizational, social, and normative implications and to situate them within broader power relations and strategic trajectories.
3 Theoretical Background
The conceptual foundation of the workshop design combines the methodical approach of CBL with the reflective logic of the original Implication Fan as a starting point to arrive at the new Implication Canvas. CBL conceptualizes learning as situated in authentic problem contexts and driven by participants’ active sense-making. We transfer this general perspective to the learning of ethical assessment, arguing that it is particularly well suited to the case-based, experience-oriented, and collectively reflective character of co-determination practices. CBL aims to encourage learners to develop solutions to complex, real-world problems by applying interdisciplinary knowledge and working collaboratively. In the present project, CBL did not function as a separate teaching module; instead, it informed the redesign of the Implication Fan into the Implication Canvas by emphasizing participant agency, collaborative inquiry, action orientation, and iterative reflection. The essential aspects of the CBL are based on (Gallagher & Savage, 2023; Nichols & Cator, 2008) as follows:
-
1.
Authentic challenges: Learners work on real-world problems that are relevant to them and their community.
-
2.
Interdisciplinarity: Knowledge and skills from different disciplines are integrated.
-
3.
Collaboration: Learners work in teams, often with external partners such as experts or community members.
-
4.
Action orientation: The focus is on the development and implementation of solutions, not just on theoretical discussion.
-
5.
Furthermore, continuous reflection on the process and the results enables a self-reflective learning process.
Those aspects are implemented in the canvas on several levels but are also caused by the surroundings. Especially the first two are part of the context. We should note that we claim that 3–5 are fostered by the design of the Implication Canvas. In the context of the works council, this simply mirrors the nature of judging concrete interventions. But also, beyond this, it forces the members to think about tangible factors of implementation, limits, and chances, which goes beyond a mere judging role. By working in a workshop setting, we get different stakeholders in one place. The goal is concrete and not just theoretical. By an iterative design, we furthermore foster reflection.Footnote 4 Even if some risks or evaluative criteria are already known to the workshop participants, it makes a substantial difference when they draw these conclusions themselves in relation to their own case. In such situations, it can be argued that a sense of agency is fostered, strengthening participants’ ability to defend their judgments in subsequent co-determination processes. Regarding the roles of the course/workshop leaders, it should be noted that, a supporting and guiding role is taken instead of imparting knowledge directly. The leader defines the framework and provides the resources to promote the learning process. CBL emphasizes the learning process itself, not just the outcome. Reflection and iterative improvement are central elements. In summary we aim to empower learners to construct knowledge actively and independently, supported by targeted guidance.
If we enable learners to take perspectives and optimize processes themselves – so we argue – socio-ethical aspects will be internalized. This refers to any technology in general, not just aspects of certain technology. It is in contrast to check-box approaches, in which “facts” are presented, so to speak, about what needs to be considered. We argue that this is how the Implication Fan can be rediscovered, especially in a form adapted to the actual processes. It is therefore central to our approach that participants articulate problems and relevant dimensions themselves, if necessary by transferring them from analogous cases. Given the increasing number of similar digitalization and AI projects across organizations, such transfer processes become more likely and more meaningful over time. The underlying hypothesis is that self-developed heuristics and evaluative factors (for example concerning marginalized groups or long-term consequences) are more robustly anchored in organizational practice than externally imposed criteria.
One aspect of achieving this is the constant practice of taking on non-own perspectives. Partly as a suggestion, partly in a structured role. This putting yourself in somebody else’s shoes is expected to bring new perspectives into the workshop (even better we would of course have direct representation of all stakeholders, which is unfortunately not realistic). This has already proven to be useful in other contexts (cf. Willis et al., 2022) one particular kind of example can be found in democratic institutions, when someone speaks for a non-present group, see the idea of ombudspersons for children (or animals, or future generations) (cf. About the Ombudsman for Children in Sweden, 2021; Animal Care International, 2025).
The developed Implication Canvas allows participants to reinvent solutions, thereby producing localized knowledge that is owned by the participants themselves and embedded in their actual organizational practices.
4 Research Design
The following section provides insight into our methodological approach, the original Implication Fan, and the different workshop settings.
4.1 Research Methodology – Exploratory Qualitative Research Design
In order to enhance the autonomous decision-making capacity of works councils, to address the challenges outlined in Sect. 2 and facilitate the co-creation of the Implication Canvas for socio-ethical assessment, the following methodological approach was pursued. To analyze, reflect upon, and further develop the Implication Fan, we conducted short-form participant observation based on an ethnographic approach following Breidenstein et al. (2015) during the facilitated workshop sessions. This form of situated observation enabled us to capture interaction dynamics, discursive negotiation processes, and implicit forms of meaning-making that occurred during the collaborative work with the Implication Fan. The researchers assumed a dual role as facilitators and participants, maintaining a balance between social proximity and analytical distance. Particular care was taken to avoid expressing personal opinions or suggestions that could bias the development of the tool. To supplement this, aspects of the co-creation approach, as described by Vargas et al. (2022), were incorporated. The works council members became active designers of the tool at various stages of development, with each member able to contribute their different areas of expertise directly and with minimal barriers.
At the end of the workshop, we handed out a two-page questionnaire to accompany the observations (see Appendix 2 for a translated version). The purpose of the questionnaire was to gather feedback regarding participants’ satisfaction with the workshops, perceived practical relevance, and remaining questions concerning the applicability and further development of both the Implication Fan and the underlying workshop design. In addition, the questionnaire had the advantage that it could be completed anonymously, thus avoiding a potential barrier to critical comments. Participant numbers varied across settings: seven participants in the first works council workshop, five in the second, and five in the third workshop. The works council members had diverse professional backgrounds and represented a variety of perspectives, with roles ranging from bus drivers to office workers, clerks and asset managers. The questionnaire served as a qualitative supplementary instrument, enriching insights gained through observation with participants’ subjective assessments and thereby providing a more differentiated understanding of user experiences with the Implication Fan. The insights generated through participant observation and the questionnaire formed the basis for the iteratively integrated adaptations to the Implication Canvas discussed in Sect. 5.2 and 5.3.
4.2 The Original Implication Fan as a Working Basis
The so-called Implication Fan (see Figure 1), developed at the Berlin Ethics Lab of TU Berlin (Fischer & Mehnert, 2021), serves as the foundational epistemic tool for this study.
In its original conception, the Implication Fan constitutes a structured canvas designed to support joint reflection and deliberation within diverse participant groups by providing a causal, temporal, and thematic scaffold for discussion. As Fischer and Mehnert describe, the Implication Fan “helps to think through possible short- and long-term implications of a technology, innovation or research result for the socio-technical system” (2021). The tool enables the systematic elicitation and organization of diverse stakeholder perspectives, while also allowing participants to articulate any additional concerns that are not predetermined or readily categorizable. Through process-oriented engagement with real or hypothetical application scenarios, participants derive multiple implications that may affect workforce segments, role statuses, professional responsibilities or organizational task distributions, for example.
More specifically, the Implication Fan consists of four interconnected components. First, it starts from a concrete application scenario that is clarified with descriptions regarding “who uses the result”, “why is it being used”, ”context of use”, and “how it is used” (middle). Second, the users of the tool describe the individual-, technical-, and systemic prerequisites for the application, with choices between temporal horizons, namely, “that requires”, “and that needs”, and “and that needs” (bottom). Third, the consequences of these levels are derived across “then happens”, “and therefore happens”, and “that leads to” (top). Fourth, the process concludes with the synthesis of central ethical issues (right) and the marking of interdependencies among identified implications.
4.3 Case Examples
The subsequent section will provide an overview of the scenarios that were utilized in the workshops with the aforementioned municipal utility provider’s works council.
In the first workshop with the works council, no real implementation project was under discussion; therefore, we used a case based on a commercial, AI-driven talent management platform designed to suggest career development and training paths within HR departments. This served as a proxy for expected future co-determination challenges.
The second workshop addressed a real case: the use of Microsoft 365 Copilot for automated meeting minutes. This scenario provided a direct and practical context for examining implications relevant to co-determination processes and workplace governance.
The third workshop focused on the automatic generation of emails using Microsoft 365 Copilot and its implications for corporate processes and stakeholders. This approach enabled the knowledge acquired from the preceding workshop to be reused and further expanded. The workshop was also used to determine whether the canvas version developed to date was sufficiently ready for everyday use. The objective of this evaluation was to ascertain whether works councils’ members would be able to work independently and satisfactorily with the tool for their future work.
5 Empirical Results
In the subsequent sections, we delineate our methodology, which entailed the structuring and configuration of the workshop to facilitate the implementation of the adaptations to the canvas. We provide concrete insights into how we addressed the challenges outlined in Sect. 2, how we integrated our pedagogical background, and which results from participant observation and the questionnaires led to which adaptations.
5.1 Workshop Design and Procedure for Applying the Implication Canvas
The following section delineates the distinctive features and the workshop approach in which the Implication Canvas was subjected to testing and refinement. The workshop design is intentionally flexible and can be adapted to different organizational and educational settings. After a brief welcome, clarification of the workshop’s objectivesFootnote 5, and an outline of the process, facilitators may choose to include a short introductory exercise to establish group spirit. Examples include inviting participants to introduce themselves and name a value they consider relevant to technological innovation, or forming small groups based on commonalities identified by the participants themselves. Such activities can help create a supportive working atmosphere but may be omitted when the group already knows one another, is too large, or when time constraints make them impractical.
Following the introduction, the application scenario is presented by a designated participant, via video material, or by the facilitators. A short clarification phase ensures a shared understanding of the scenario. The group then completes the scenario description on page two of the canvas. Participants are reminded to incorporate the user and stakeholder groups listed in the “Who uses the outcome?” field into their subsequent reflections, enabling perspective-taking practices without requiring formal subgroup formation. Depending on the workshop duration, at least one break is recommended (see Appendix 1).
We experimented with different formats for working with the canvas, ranging from collective brainstorming to silent group work followed by short presentations. The choice of method depends on group size, dynamics, facilitation style, and scenario complexity. In a test environment conducted in a teaching context in which two different scenarios were worked on simultaneously, we relied on group work with brief presentations on prerequisites, consequences, and solutions. In the first workshop, we assigned the task of assuming additional roles and tested the effectiveness of this approach. However, this proved to be unnecessarily complicated. In the next iteration, other perspectives were incorporated into the elaborations and discussions, even without separately assigned roles. We designed the work with the canvas at the beginning in a joint brainstorming session of the first three fans of the prerequisites to ensure that the approach was understandable. Then, the consequences and solution approaches were filled out independently, each followed by an explanation and discussion. In the second workshop with the works council, the canvas was completed through collective brainstorming due to a significantly shortened time schedule. For the final iteration of the third workshop with the works council, we adopted a two-group method: one group focused on prerequisites and the other on consequences. After completing their sections independently, the groups exchanged their worksheets and supplemented each other’s work in silent reflection. The results were then jointly discussed, followed by the identification of interdependencies, which informed the subsequent development of solution pathways. This approach reduces conformity pressures, encourages diverse viewpoints, and showed indications for deeper individual engagement with the material.
5.2 Reflections and Adaptations Leading to the Implication Canvas
For the reader’s convenience, the central challenges from Sect. 2 are listed below:
-
a.
Epistemic differences between management and works councils.
-
b.
Decision-making under time pressure and uncertainty.
-
c.
The assessment of socio-technical implications of AI systems.
-
d.
Prestructure decision-making processes and render alternative technological pathways less visible.
-
e.
The protection of collective labor interests under transformation conditions.
The original Implication Fan’s design and structure facilitate examination of multiple levels of analysis and offers a higher degree of abstraction regarding the prerequisites and consequences of innovative technologies. For these reasons, this tool has been chosen as the basis for our adaptations. Moreover, due to the structure and orientation of the tool, the possibility was recognized to substantially address the key challenge a. of minimizing epistemic differences by linking experience-based knowledge and e., through comprehensive risk assessments for different levels. The focus of the adaptation was on guiding and facilitating the formation of a qualified opinion regarding the socio-ethical implications of AI systems (c.) for works councils in co-determination processes. Furthermore, our goal was to develop the Implication Fan into a flexible instrument that could be used in different companies, allowing works councils to configure both the tool and the workshop concept according to their specific needs. Initially developed for research projects and academic settings, the Implication Fan was subsequently adapted to align with the requirements of works council projects, demonstrating its adaptability and versatility. Therefore, in order to achieve greater accessibility for participants with diverse professional backgrounds (e.), possible adaptations were envisaged. The ensuing sections will present the results and the outcomes of the design adaptations that informed the development of the new Implication Canvas.
5.3 Design Modifications
The design modifications underwent three major iterations, in addition to several minor refinements, all of which were based on feedback from observations, questionnaires and informal discussions. The first adaptation phase, which took place before the first workshop, focused on making the tool applicable in a co-determination setting. This included translating the canvas into German, restructuring it into a numbered sequence, and reformulating key terminology and guiding questions. “Ethical issues” for example, were replaced with “Solution Approaches”, aligning the exercise more closely with a CBL logic that emphasizes practical problem-solving. Moreover, a solution-oriented structure confers the advantage of enhancing practical relevance and enabling users to formulate concrete proposals on how best to address potential misuse and use (d.). Nevertheless, ethical issues have not been neglected and have repeatedly been addressed in the specific canvas elaboration.
An “Open Aspects” field was added to capture insights that did not fit predefined categories, and the upper task description was simplified to focus with wording on the identification of prerequisites, consequences, and potential solutions for a technological innovation. These changes aimed to shift the framing from abstract reflection to actionable deliberation. The explicit reference to actual technical practical projects enables the proactive avoidance of potential risks as they emerge (d.). This prototype (see Fig. 2) was tested in a university setting to identify and address potential issues before being refined for the works council kickoff workshop.
5.3.1 First Workshop 22.05.2025
After the first works council workshop, several participants described the tool as useful but still overly complex. Survey results reflected a mixed assessment: three participants found the fan difficult to navigate, while four rated it as reasonably accessible. As a result, further adjustments were introduced. The “Open Aspects” section was removed to create more space and streamline the layout, as the entries were not immediately relevant and were therefore recorded elsewhere. To increase available space for preconditions and consequences and allowing greater flexibility in preparing or distributing case materials, the case description was moved to a separate page. The structure of the temporal levels was simplified by reducing them to two tiers: “this requires/needs” and “then this happens/which leads to”. It is posited that this will serve to minimize redundancy, improve clarity, and assist in reducing the challenge of insufficient time required to utilize the tool (b.). Terminology was further simplified (e.g., replacing “stakeholders” with “persons and groups”), because there was some ambiguity surrounding these terms. Furthermore, the color scheme was aligned with ver.di’s (United Services Trade Union) corporate design, ensuring better readability and visual coherence. The identical color scheme is intended to convey ver.di’s ideals of solidarity, freedom, democracy, and justice (ver.di, 2025), thereby encouraging open discussion (e.). As we intended to provide additional input on the ethical considerations of LLMs for the second workshop, as well as the canvas moderation, it was necessary to impose a stricter timeframe. To accomplish this, the application scenario was prefilled, and a concise exposition on the particularities of automated meeting minutes was provided. A ten-minute supplementary phase on the completed aspects was then conducted. In doing so, we took care to avoid introducing our own biases into the subsequent discussion by presenting case studies and aspects of AI applications in a neutral manner that did not directly correspond to the meeting’s theme. For technical implementations with a proven foundation, it was proposed that the application scenario could be completed in less time than a conventional works council meeting (b.). For the second version see Fig. 3.
5.3.2 Second Workshop 20.10.2025
A second workshop with the works council in October 2025 confirmed that usability had improved: two of five participants reported using the adapted version as very well, and the remaining three rated it as rather well. Nevertheless, both spoken feedback and observation indicated that the fan-shaped structure was still somewhat cumbersome. Therefore, the overall structure was transformed from a “fan” into a grid of square boxes. This layout better aligns with common matrix-based visualization formats, making the tool more intuitive and easier to use in both physical and digital formats. The preconditions were positioned at the top and the consequences at the bottom to reflect sequential logic. The arrows were redesigned to free up space and emphasize the causal linkages between preconditions and consequences in the center, thereby supporting the identification of interdependencies (d., e.). To reflect this redesign, the tool was renamed Implication Canvas. Minor textual and visual adjustments were made to enable the canvas to be self-explanatory from the outset of the workshop. These included bolding key elements and moving the scenario to page two, so that the task description could be read at a glance. In addition, this should further reduce the training period for new users (b.). As part of this change, the title of the scenario was changed from “The Future Application” to “Application Scenario”, since not only future applications should be reflected upon, but also already established ones. The emphasis on potential solutions proved to be of notable benefit during the workshops, as the participants repeatedly encountered possible solutions while deliberating on the prerequisites and consequences. These solutions could then be recorded immediately and subsequently elaborated upon in greater depth (d., e.). For the final version of the canvas see Fig. 4 (German version) and Fig. 5 (English version).
5.3.3 Third Workshop 15.12.2025
The third workshop served as a trial run to determine whether the workshop’s format and the adapted canvas were sufficiently developed to meet the needs of the works councils. During this workshop, the researchers refrained from intervening, focusing primarily on moderation.
The results of the questionnaire show that three out of five participants rated the final version as “very well”, and the remaining two rated it as “rather well”. As free-text feedback, it was added by one person that the method fits their future co-determination work. While discussions about solution strategies were praised, one person criticized the excessive focus on “what if” scenarios. Such considerations are probably unavoidable, especially since they can be important for potential technological impact assessments. However, the practical focus should probably always be paramount for works councils’ activities in order to adequately protect collective labor interests under transformation conditions (e.). It was also suggested that the canvas be made available as an editable digital version, reflecting the initial plan and its implementation. The observations revealed no major issues. However, the moderator must be mindful of the time frame, as discussions can otherwise become too lengthy. Finding the balance between adhering to the time frame and not prematurely interrupting discussions appeared essential.
6 Analysis of the Outcomes in Relation to Three Pressing Challenges Identified
The workshop results are particularly informative with regard to three challenges identified in Sect. 2: epistemic asymmetries between management and works councils (a.), the assessment of socio-technical implications of AI systems (c.), and the protection of collective labor interests under transformation conditions (e.). These three challenges were selected for closer analysis because they recurred throughout the workshop series, gained empirical salience across the iterative refinements of the canvas, and bear directly on the question of how works councils can be supported in developing structured, practice-oriented, and socio-ethically informed judgments about AI-related workplace transformations under conditions of uncertainty and asymmetrical expertise.
Regarding epistemic asymmetries (a.), the workshops did not entirely eliminate differences in technical expertise. However, they reconfigured how such differences entered the deliberative process. The canvas created a shared structure in which technical assumptions, organizational and individual preconditions, and anticipated consequences had to be articulated in a form accessible to participants with heterogeneous backgrounds. This shifted the interaction from a situation in which specialized knowledge could function as a largely unchallengeable source of authority toward one in which claims became discussable in relation to workplace experience, legal concerns, and organizational consequences. This shift can be seen as important not only at the level of discussion quality but also concerning agency in co-determination practice: From emphatic statements on transformative effects on the individual employee, towards a view of the structural and organizational conditions, consequences, and—ultimately—solution approaches, members of the works council were incentivized to adopt a holistic, but structured perspective on ways to act and engage with technologically innovative proposals. A more differentiated understanding of AI systems and their organizational implications can enable works councils to formulate more specific demands, including the more needs-based involvement of external experts to which they are legally entitled in Germany. In this sense, the canvas does not merely support reflection; it may also strengthen the capacity to assert procedural rights in a more informed and self-determined manner. Rather than responding with only general skepticism or surface-level rejection, participants were increasingly able to articulate targeted questions, concrete conditions, and negotiable requirements. This may reduce the scope and works council’s susceptibility for strategic manipulation through informational asymmetries and contribute, at least partially, to a rebalancing of power relations. At the same time, these observations should be interpreted with caution. Since the workshop series did not extend into a longitudinal observation of subsequent works council meetings, the longer-term organizational effects of this enhanced reflective capacity can only be inferred indirectly at this stage. The present findings therefore indicate an increased potential for agency rather than a fully demonstrable institutional transformation.
The workshops also yielded first insights regarding the assessment of socio-technical implications of AI systems (c.). This became visible not only in the range of issues raised, but also in the increasing depth and differentiation of the discussions, especially in the later workshop iterations. The second and third workshops, both of which focused on Microsoft 365 Copilot use cases, illustrated that participants moved well beyond a narrow assessment of technical functionality and engaged more systematically with the social and normative dimensions of AI use in workplace settings. In the second workshop on automated meeting minutes, the initial appeal of AI-supported transcription and summarization was still clearly present. However, this relatively enthusiastic orientation became noticeably more cautious once participants reflected on the error-proneness of generative AI and on the complexity of human communication. Participants themselves raised concerns that ironic or mocking remarks could be misinterpreted, that subtle communicative nuances might be lost, and that interpersonal exchange could be adversely affected when machine-generated records begin to shape workplace communication. Importantly, these concerns were not simply introduced by the facilitators but emerged from the works council’s own deliberations. This suggests that the canvas helped participants generate socio-technical critique from within their situated understanding of organizational processes. A comparable pattern emerged in the third workshop on the automatic generation of emails. Here, participants discussed questions of transparency regarding AI use, insisted that responsibility for generated outputs must remain with human actors, and emphasized the necessity of reviewing AI-generated text before circulation. In addition, risks of deskilling were explicitly named and discussed. Workshop participants anticipated how AI-supported process transformations can spur a devaluation of educational backgrounds and expertise. These workshop findings resonate with broader developments in the German context. Survey evidence indicates that some companies are considering replacing highly qualified employees with less-qualified workers who operate AI tools (Ruffert, 2026). This illustrates how AI-supported process changes can entail layoffs, deskilling, and shifts in responsibility. At the same time, the deliberations did not remain one-sidedly critical. Participants also reflected on possible benefits, such as time savings, the reduction of routine burdens, and more efficient processes that might, under certain conditions, create more room for social interaction. Precisely this ambivalence can be considered analytically valuable: it shows that the workshops did not simply elicit approval or rejection but fostered a more nuanced socio-ethical engagement with competing organizational consequences.
A third insight concerns the protection of collective labor interests under transformation conditions (e.). Here, the workshops demonstrated that many relevant risks of AI adoption do not appear as immediate harms at the moment of implementation. Instead, they emerge through indirect chains of consequences, for example via subtle shifts in responsibility, qualification requirements, or expectations of constant availability and efficiency. By structuring reflection around prerequisites, consequences, interdependencies, and responses, the canvas enabled participants to articulate these indirect effects earlier and in a more systematic manner. This can be particularly important for co-determination practice, since collective interests can only be defended effectively when emerging risks become visible before they are sedimented in routine organizational processes. Supporting claims about emerging risks with plausible, documented, and structured causal chains may not prevent a company from introducing a disruptive innovation in the workplace. However, it may lead to a more cautious implementation process that takes risks seriously and keeps options for retrievability and careful reversion of (parts of) the changes in mind.
Taken together, these findings suggest that the Implication Canvas is most valuable not because it produces definitive judgments, but because it creates a procedural setting in which socio-technical claims become discussable across different forms of expertise. Its contribution lies less in replacing technical knowledge than in organizing a deliberative process through which legal, experiential, and ethical perspectives can foster future co-determination practices.
7 Discussion and Conclusion
Overall, the iterative adaptations demonstrate tangible outputs of a co-creation process that responded directly to user feedback and aimed at balancing analytical rigor with practical accessibility (c.). The resulting Implication Canvas preserves the fundamental deliberative idea of the original Implication Fan, while improving its usability and inclusivity for diverse actors (e.), particularly for non-academic participants. More specifically, the structure and task descriptions have been streamlined, rid of academic jargon, and clarified (b.). By transforming abstract socio-ethical reflection into a structured yet open framework for collaborative exploration, the tool’s adaptations were made to support works councils in identifying assumptions, tracing consequences, and generating context-sensitive strategies in the governance of AI and digital transformation (a, c, d, e). Consequently, there is a possibility that working with the canvas can contribute to a more humane business reality in the long term. Therefore, we posit that the tool contributes as a concrete means to mitigate the central challenges of co-determination in a fast-paced digital transformation.
Moreover, the iterative co-creation process with council members further aligns with CBL-inspired and participatory design approaches to qualitative tool development, in which instruments are not only evaluated but progressively redesigned to balance analytical depth with practical usability. By demonstrating how a CBL-informed reflection tool can be embedded in ongoing co-determination processes around AI, the study advances research on workplace learning and socio-ethical evaluation instruments and underscores the relevance of challenge-based frameworks for strengthening institutional capacities for responsible digital transformation.
As generative AI systems become embedded in organizations and organizational decision-making, works councils are increasingly required to engage with technologies whose implications reach beyond immediate operational contexts, and require a structural perspective in line with current trends in the AI ethics debate (cf. Bolte & Van Wynsberghe, 2025). Rapid transformation processes underline the need for strengthening digital sovereignty as a core competence. Here, strengthening works councils’ reflective capabilities can both be regarded as a source for supporting the digital sovereignty of enterprises as a whole, through more effective co-determination, while—at the same time—increasing the digital sovereignty of the works council itself also alleviates epistemic asymmetries between management, technical experts, and worker representatives. In this sense, digital sovereignty extends beyond technical literacy to encompass the ability to interpret, evaluate, and influence technological developments within their social and normative contexts. The workshop series suggests that such sovereignty is not merely a matter of knowledge acquisition, but also of procedural agency: participants increasingly moved from broad concern or general skepticism toward more differentiated questions, concrete conditions, and negotiable requirements. This indicates that the canvas may support a more self-determined exercise of co-determination rights, including the more informed and needs-based involvement of external expertise where appropriate. We suggest that the approach presented in this contribution transcends the dichotomous separation of individual and national/territorial digital sovereignty (e.g., Couture et al., 2024), as it works on the assumption that structures of individuals, such as the works councils, may be an appropriate target of interventions to support digital sovereignty. In a certain sense, their abilities for effective, reflected, and informed co-determination—or rather their lack thereof—can be regarded as a residual problem resulting from companies rushing to maintain their influence and agency against the backdrop of the digital transformation (Herzog et al., 2025).
The iterative development of the Implication Canvas demonstrates how participatory design methods can operationalize the ambition to structurally support digital sovereignty, in the above-mentioned sense. By translating abstract reflection into a structured yet flexible process, the canvas provides a collaborative framework that enables diverse actors to articulate assumptions, visualize potential consequences, and deliberate alternative courses of action. Its adaptability across different organizational and educational settings underscores its broader relevance in contexts where socio-technical complexity demands collective judgment. Although we have tested different workshop approaches, it remains uncertain whether the modularity idea within the concept will be used for future works council deliberations. However, it has been demonstrated that the latest version of the concept and tool creates a useful deliberative space for various applications and can therefore be effective in practice. More instantiations of collective deliberations using the canvas would surely help to elucidate remaining shortcomings or potentials. Since the works council was already familiar with the tool, using it was self-explanatory after the first instance. It is unclear whether this would also be the case for works councils’ members without prior knowledge of the canvas. Nevertheless, it should be noted that one person new to the tool had no difficulty using it during the last workshop. At the same time, the present study does not permit firm conclusions about longer-term organizational effects, since subsequent works council deliberations after the workshops were not observed. The findings therefore point to an increased potential for informed and agency-enhancing co-determination, while the extent of its sustained institutional uptake remains an open question.
Future research should focus on longitudinal studies of the tool’s integration into organizational decision-making processes, as well as on comparative analyses across different industries and other international co-determination settings. It would also be valuable to examine more systematically how such participatory reflection tools relate to existing literature on qualitative tool development and co-creative methodology, especially regarding the relation between usability, iterative redesign, and epistemic robustness in practice. Such work could deepen understanding of how participatory reflection tools contribute to institutional learning and to the sustainable governance of AI in the workplace. Ultimately, enhancing works councils’ capacity for informed, reflective, and sovereign co-determination remains a decisive factor for ensuring that digital transformation aligns with democratic and socially responsible principles.
Data Availability
For insights into the anonymized questionnaire data, please contact the corresponding author: robin.preiss@web.de.
Code Availability
Not applicable.
Notes
The project was funded by the Hans-Böckler-Foundation in the “Transformation” funding line.
For example, a black-box model is a machine learning model whose internal structure and decision logic are not directly visible or interpretable for users, while it generates predictions based on input data (Hassija et al., 2024, p. 47).
And even worse, an AI-implementation plan based on uninformed management structures or those ignorant of actual needs, for instance to boost other metrics, like market capitalization would even more make the case for the need of a strong works council’s involvement.
We note that the impact of such reflective processes can be found in other pedagogical traditions as well. In particular, we mention the RME Tradition (Realistic Mathematical Education), in which learners re-invent knowledge, cf. Freudenthal (1973).
It is also beneficial to send the objectives and a rough outline of the workshop day to all participants by email at least one day in advance, to give them time to familiarize themselves with the content and prepare mentally for the workshop.
References
About the Ombudsman for Children in Sweden. (2021). Barnombudsmannen. https://www.barnombudsmannen.se/english
Aktien Check (2026, June 14). AI Reshapes the Workforce: German Minister Sees Every Job Touched by 2030 as Firms Eye Degree-Free Replacements. (worldwide). Aktien Check. https://www.aktiencheck.de/news/Artikel-AI_Reshapes_the_Workforce_German_Minister_Sees_Every_Job_Touched_by_2030_as_Firms_Eye_Degree_Free_Replacements-19848059
Animal Care International. (2025). Tierschutz-Ombudsmann: Tierschutz-Ombudsmann Aufgaben und Kontaktadressen in ganz Österreich. Animal Care International: Menschen Für Tiere. https://www.animalcare-international.org/tierschutz-ombudsmann
Bader, V., & Kaiser, S. (Eds.). (2020). Arbeit in der Data Society – Perspektiven auf die Zukunft von Mitbestimmung und Personalmanagement. In Arbeit in der Data Society: Zukunftsvisionen für Mitbestimmung und Personalmanagement (pp. 3–14). Springer Fachmedien Wiesbaden. https://doi.org/10.1007/978-3-658-32276-2
Bainbridge, L. (1983). Ironies of automation. Automatica, 19(6), 775–779. https://doi.org/10.1016/0005-1098(83)90046-8
Bolte, L., & Van Wynsberghe, A. (2025). Sustainable AI and the third wave of AI ethics: A structural turn. AI and Ethics, 5(2), 1733–1742. https://doi.org/10.1007/s43681-024-00522-6
Bombaerts, G. (2021). Challenge-based learning to improve the quality of engineering ethics education. In H.-U. Heiß, H.-M. Järvinen, A. Mayer, & A. Schulz (Eds.), Proceedings—SEFI 49th Annual Conference: Blended Learning in Engineering Education: Challenging, Enlightening -and Lasting? (pp. 1574–1581). European Society for Engineering Education. SEFI Annual Conference.
Breidenstein, G., Hirschauer, S., Kalthoff, H., & Nieswand, B. (2015). Ethnografie: Die Praxis der Feldforschung. UTB GmbH.
Butollo, F., Gerber, C., Görnemann, E., Greminger, L., Katzinski, A., Kulla, M., Sirman-Winkler, M., & Spott, J. (2024). Die Symbiose von generativer KI und Arbeit: Erweiterung der Horizonte oder Erosion menschlicher Kompetenz? Weizenbaum Discussion Paper, (41).
Couture, S., Toupin, S., & Mayoral Baños, A. (2024). Resisting and Claiming Digital Sovereignty: The Cases of Civil Society and Indigenous Groups. Policy & Internet, 16(4), 739–749. https://doi.org/10.1002/poi3.434
Dindler, C., Iversen, O. S., Bossen, C., Kaspersen, M. H., & Dalsgaard, P. (2026). Enacting Utopias: A Participatory Design Approach to AI. Proceedings of the 19th Participatory Design Conference 2026 1: Full Papers, 118–128. https://doi.org/10.1145/3796624.3796657
Doulougeri, K., Vermunt, J. D., Bombaerts, G., & Bots, M. (2024). Challenge-based learning implementation in engineering education: A systematic literature review. Journal of Engineering Education, 113(4), 1076–1106. https://doi.org/10.1002/jee.20588
Ferdman, A. (2025). AI deskilling is a structural problem. AI & SOCIETY. https://doi.org/10.1007/s00146-025-02686-z
Fischer, N., & Mehnert, W. (2021). ETHICS TOOLBOX: Implication Fan. Technische Universität. https://www.tu.berlin/en/philtech/berlin-ethics-lab/ethics-toolbox Ethics Lab.
Freudenthal, H. (1973). Mathematics as an educational task. Springer Netherlands. https://doi.org/10.1007/978-94-010-2903-2
Frey, C. B., & Osborne, M. A. (2017). The future of employment: How susceptible are jobs to computerisation? Technological Forecasting and Social Change, 114, 254–280. https://doi.org/10.1016/j.techfore.2016.08.019
Gallagher, S. E., & Savage, T. (2023). Challenge-based learning in higher education: An exploratory literature review. Teaching in Higher Education, 28(6), 1135–1157. https://doi.org/10.1080/13562517.2020.1863354
Gerbracht, M., Krüger, M., De Castro Leal, D., Tolmie, P., & Wulf, V. (2024). We are no Luddites!—CSCW, Co-Determination and Digital Transformation in Germany. Proceedings of the ACM on Human-Computer Interaction, 8(CSCW1), 1–23. https://doi.org/10.1145/3637430
Gesetz zur Förderung der Betriebsratswahlen und der Betriebsratsarbeit in einer digitalen Arbeitswelt (Betriebsrätemodernisierungsgesetz), 32 Bundesgesetzblatt Jahrgang 2021 § 1 (2021). https://www.bgbl.de/xaver/bgbl/start.xav?startbk=Bundesanzeiger_BGBl&jumpTo=bgbl121s1762.pdf#/text/bgbl121s1762.pdf?_ts=1765282710070
Grasy, J., Seibold, B., & Klengel, E. (2024). KI und algorithmische Systeme verstehen, bewerten und begrenzen—So gelingt Mitbestimmung bei KI; Praxisbeispiel und Impulse. Mitbestimmungspraxis, (59). https://www.imu-boeckler.de/de/faust-detail.htm?sync_id=HBS-008807
Guliyev, H. (2023). Artificial intelligence and unemployment in high-tech developed countries: New insights from dynamic panel data model. Research in Globalization, 7, 100140. https://doi.org/10.1016/j.resglo.2023.100140
Haipeter, T., Wannöffel, M., Daus, J. T., & Schaffarczik, S. (2024). Human-centered AI through employee participation. Frontiers in Artificial Intelligence, 7, 1272102. https://doi.org/10.3389/frai.2024.1272102
Hassija, V., Chamola, V., Mahapatra, A., Singal, A., Goel, D., Huang, K., Scardapane, S., Spinelli, I., Mahmud, M., & Hussain, A. (2024). Interpreting Black-Box Models: A Review on Explainable Artificial Intelligence. Cognitive Computation, 16(1), 45–74. https://doi.org/10.1007/s12559-023-10179-8
Herzog, C., Breyer, S., Leinweber, N. A., Preiß, R., Sonar, A., & Bombaerts, G. (2022). Everything you want to know and never dared to ask. A practical approach to employing challenge-based learning in engineering ethics. Towards a New Future in Engineering Education New Scenarios That European Alliances of Tech Universities Open Up, 1224–1232. https://doi.org/10.5821/conference-9788412322262.1392
Herzog, C., Zetti, D., & Preiß, R. (2025). Digital sovereignty as an ill-structured (or wicked?) problem. Uncertain Journeys into Digital Futures: Inter- and Transdisciplinary Research for Mitigating Wicked Societal and Environmental Problems, 1, 191–206. https://doi.org/10.5771/9783748947585-191
Kellogg, K. C., Valentine, M. A., & Christin, A. (2020). Algorithms at Work: The New Contested Terrain of Control. Academy of Management Annals, 14(1), 366–410. https://doi.org/10.5465/annals.2018.0174
Koch, M., & Lodefalk, M. (2025). Artificial Intelligence and Worker Stress: Evidence from Germany. Digital Society, 4(1), 5. https://doi.org/10.1007/s44206-025-00160-3
Krook, J. (2025). When autonomy breaks: The hidden existential risk of AI. AI & SOCIETY, 40(8), 6011–6024. https://doi.org/10.1007/s00146-025-02397-5
Krzywdzinski, M. (2024). Zwei Welten der KI in der Arbeitswelt: Wie Management und Betriebsräte die Einführung und Nutzung von KI-Anwendungen gestalten. Weizenbaum Discussion Paper, (39).
Krzywdzinski, M., Gerst, D., & Butollo, F. (2023). Promoting human-centred AI in the workplace. Trade unions and their strategies for regulating the use of AI in Germany. Transfer: European Review of Labour and Research, 29(1), 53–70. https://doi.org/10.1177/10242589221142273
Lewis, D., Reijers, W., & Pandit, H. (2017). Ethics Canvas Manual. https://www.ethicscanvas.org
Malmqvist, J., Rådberg, K. K., & Lundqvist, U. (2015, June 8). Comparative analysis of challenge-based learning experiences. Proceedings of the 11th International CDIO Conference.
Milanez, A. (2023). The Impact of AI on the Workplace: Evidence from OECD Case Studies of AI Implementation (OECD Social, Employment and Migration Working Papers No. 289; OECD Social, Employment and Migration Working Papers, Vol. 289). https://doi.org/10.1787/2247ce58-en
Milanez, A. (2025). Exploring win-win outcomes of algorithmic management: Lessons from a laboratory experiment on worker consultation (43rd ed., OECD Artificial Intelligence Papers 43) [OECD Artificial Intelligence Papers]. OECD Publishing. https://doi.org/10.1787/84b59397-en
Moore, P. V. (2019). OSH and the Future of Work: Benefits and Risks of Artificial Intelligence Tools in Workplaces. In G. Vincent, & Duffy (Eds.), Digital Human Modeling and Applications in Health, Safety, Ergonomics and Risk Management. Human Body and Motion, 11581, 292–315. Springer International Publishing.
Nichols, M. H., & Cator, K. (2008). Challenge Based Learning, A White Paper. Apple, Inc.
Niewerth, C., Massolle, J., Schaffarczik, S., & Grabski, C. (2022). Betriebsräte in der doppelten Transformation: Ein Transferforschungsprojekt zur Organisationsentwicklung von Betriebsgremien. Hans-Böckler-Stiftung.
ODISSEI. (n.d.). Ethical Dimensions of Computational Social Science Research: A Primer. ODISSEI – Open Data Infrastructure for Social Science and Economic Innovations. Retrieved July 6 (2026). from https://odissei-data.nl/ethical-social-science/
Portegies, T., Hajri, O., Willemsen, L., & Harbers, M. (2025). Decoding AI ethics: What do ethical tools tell us about ethics? Ethics and Information Technology, 27(4), 65. https://doi.org/10.1007/s10676-025-09860-3
Ruess, P., Staffa, A., Kreutz, A., Busch, C., Gayoso, S., C. O., & Pollmann, K. (2024). Künstliche Intelligenz in betrieblichen Prozessen: Ein Vorgehensmodell zur partizipativen Gestaltung von KI-Anwendungen. HMD Praxis der Wirtschaftsinformatik, 61(2), 485–502. https://doi.org/10.1365/s40702-024-01049-4
Ruffert, A. (2026, June 12). First Companies in Germany View Artificial Intelligence as an Alternative to Qualifications and Professional Experience. ifo Business Survey. Ifo Institute. https://www.ifo.de/en/facts/2026-06-12/first-companies-germany-view-artificial-intelligence-alternative-qualifications
Stubbe, J. (2017). Von digitaler zu soziodigitaler Souveränität. In Volker Wittpahl (Ed.), Digitale Souveränität: Bürger | Unternehmen | Staat (pp. 43–59). Springer. https://doi.org/10.1007/978-3-662-55796-9
Vargas, C., Whelan, J., Brimblecombe, J., & Allender, S. (2022). Co-creation, co-design, co-production for public health – a perspective on definitions and distinctions. Public Health Research and Practice, 32(2), e3222211. https://doi.org/10.17061/phrp3222211
Véliz, C. (2024). Losing Skills. In D. Edmonds (Ed.), AI morality (pp. 91–97). Oxford University Press. https://doi.org/10.1093/oso/9780198876434.001.0001
ver.di. (2025, September 30). Idee & Tradition. Vereinte Dienstleistungsgewerkschaft. https://www.verdi.de/idee-tradition
Wang, K. H., & Lu, W. C. (2025). AI-induced job impact: Complementary or substitution? Empirical insights and sustainable technology considerations. Sustainable Technology and Entrepreneurship, 4(1), 100085. https://doi.org/10.1016/j.stae.2024.100085
Widuckel, W. (2020). Arbeit 4.0 und Transformation der Mitbestimmung. In V. Bader & S. Kaiser (Eds.), Arbeit in der Data Society: Zukunftsvisionen für Mitbestimmung und Personalmanagement (pp. 17–34). Springer Fachmedien Wiesbaden. https://doi.org/10.1007/978-3-658-32276-2
Willis, R., Curato, N., & Smith, G. (2022). Deliberative democracy and the climate crisis. WIREs Climate Change, 13(2), e759. https://doi.org/10.1002/wcc.759
Funding
Open Access funding enabled and organized by Projekt DEAL. The research leading to these results received funding from the Hans-Böckler-Foundation under Grant Agreement No 2025-953-7.
Author information
Authors and Affiliations
Contributions
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Robin Preiß. The first draft of the manuscript was written by Robin Preiß and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Ethical Approval
This study was conducted in accordance with the principles of good scientific practice. All participants provided informed consent prior to data collection. The research exclusively utilized fully anonymized data, and no personally identifiable information was collected or processed at any stage. Ethical guidelines for qualitative research and data protection were strictly followed throughout the study.
Consent for Publication
All authors consent to the submission of this paper.
Competing Interests
The authors have no financial or proprietary interests in any material discussed in this article.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Preiß, R., Sarikaya, D. & Herzog, C. Strengthening Co-Determination in Digital Transformation: A Qualitative Approach to Enhancing Works Councils’ AI Literacy and Capability for Socio-Ethical Reflection. Digit. Soc. 5, 50 (2026). https://doi.org/10.1007/s44206-026-00287-x
Received:
Accepted:
Published:
Version of record:
DOI: https://doi.org/10.1007/s44206-026-00287-x