Behind the Scenes – Research & Further Study

The exhibition brings topics relating to AI and work to life. On this page, you will find the academic studies, concepts and further resources that form the basis of the individual exhibits.

Navigation:
Station 1: What is artificial intelligence?
Station 2: Humans and AI working as a team
Station 3: Focusing on people in the changing landscape of occupations within professional fields
Station 4: The importance of job identity
Station 5: Skills shortage – AI safeguards knowledge
Station 6: Changes in the regional employment structure
Station 7: Ethics and public participation in the use of AI
Station 8: HUMAINE – People with responsibility for AI

Station 1: What is artificial intelligence?

Artificial intelligence (AI) refers to a range of technical processes that process large amounts of data in order to recognise patterns The data processed includes, for example, numbers, text, photos and voice recordings. Using the recognised pattern, the AI presents a solution that would originally have required human intelligence. The AI highlights discrepancies, makes recommendations, makes predictions, generates images, and writes text and programmes.

Station 2: Humans and AI working as a team

AI can detect anomalies in images that are barely visible to the human eye – and it does so quickly and without tiring. This is why AI is used, amongst other things, to analyse X-ray images. However, as it lacks an understanding of context, AI can make incorrect decisions in difficult cases. The best solution therefore often lies in collaboration between humans and AI. Research shows how this teamwork can be structured to deliver sound and responsible results.

Authors:
Sophie Berretta, Alina Tausch, Greta Ontrup, Björn Gilles, Corinna Peifer, Annette Kluge

Abstract:
Introduction: With the advancement of technology and the increasing utilization of AI, the nature of human work is evolving, requiring individuals to collaborate not only with other humans but also with AI technologies to accomplish complex goals. This requires a shift in perspective from technology-driven questions to a human-centered research and design agenda putting people and evolving teams in the center of attention. A socio-technical approach is needed to view AI as more than just a technological tool, but as a team member, leading to the emergence of human-AI teaming (HAIT). In this new form of work, humans and AI synergistically combine their respective capabilities to accomplish shared goals.
Methods: The aim of our work is to uncover current research streams on HAIT and derive a unified understanding of the construct through a bibliometric network analysis, a scoping review and synthetization of a definition from a socio-technical point of view. In addition, antecedents and outcomes examined in the literature are extracted to guide future research in this field.
Results: Through network analysis, five clusters with different research focuses on HAIT were identified. These clusters revolve around (1) human and (2) task-dependent variables, (3) AI explainability, (4) AI-driven robotic systems, and (5) the effects of AI performance on human perception. Despite these diverse research focuses, the current body of literature is predominantly driven by a technology-centric and engineering perspective, with no consistent definition or terminology of HAIT emerging to date.
Discussion: We propose a unifying definition combining a human-centered and team-oriented perspective as well as summarize what is still needed in future research regarding HAIT. Thus, this work contributes to support the idea of the Frontiers Research Topic of a theoretical and conceptual basis for human work with AI systems.

Citation:
Berretta, S., Tausch, A., Ontrup, G., Gilles, B., Peifer, C., & Kluge, A. (2023). Defining human-AI teaming the human-centered way: A scoping review and network analysis. Frontiers in Artificial Intelligence, 6, 1250725.

Authors:
Sophie Berretta, Alina Tausch, Florian Bülow, Bernd Kuhlenkötter, Maximilian Topp, Christian Els, Corinna Peifer, Annette Kluge

Abstract:
The complementary integration of artificial intelligence (AI) in the workplace requires balancing performance goals with psychological needs, as both are essential for sustained outcomes. This study examines different workflows (AI-first and human-first) as cognitive forcing strategies to test whether they enhance performance and psychological outcomes compared to human-only and AI-only processing. In a one-factorial between-subjects experiment (N = 101) within a visual inspection task, evaluated at up to three measurement points, performance variables (accuracy, speed, error rates) and psychological variables (vigilance, flow, teaming experience, wellbeing when working with the AI) were assessed. Human-AI collaboration outperformed AI-only in error rates (η2 = 0.29) and human-only in speed (η2 = 0.11 – 0.14), but only when AI preceded human processing. The AI-first workflow enhanced teaming perception compared to human-only processing (η2 = 0.07). Moreover, human-AI collaborative processing reduced flow decrease compared to human-only processing (η2 = 0.07). Overall, AI processing preceding human processing produces the best balance between performance and psychological outcomes in safety-critical inspection tasks, supporting a holistic view of AI integration in the workplace.

Citation:
Berretta, S., Tausch, A., Bülow, F., Kuhlenkötter, B., Topp, M., Els, C., Peifer, C., & Kluge, A. (2025). Human or AI first? A holistic perspective on the sequential order of joint human-AI inspection workflows. Applied Ergonomics, 132, 104669.

Authors:
Sophie Berretta, Alina Tausch, Paul Glogowski

Abstract:
Introduction: In collaborative industrial work systems, the locus of authority—whether control over system dynamics is initiated by the system (adaptive) or by the human operator (adaptable)—can shape work experience and perceptions of the robotic partner. This study exploratively investigates how these control schemes and a static work system influence key psychological factors and which one should be favored in collaborative assembly tasks.
Methods: In an experimental laboratory study with n = 27 participants, a collaborative gearbox assembly task with a robot is used to compare adaptive and adaptable control schemes against a non-adjustable baseline. In the adaptive condition, the robot’s speed automatically adjusted to human proximity; in the adaptable condition, speed is manually adjustable via interface buttons; in the baseline condition, speed remains static. The primary endpoint was the longitudinal, comparative investigation of flow experience and perceived task demands as they represent central indicators of employees’ psychological experience in dynamic work systems. Dependent variables included additionally autonomy perception and workplace fit (task perception), and trust, safety, robot’s intelligence, and collaboration satisfaction (robot interaction perception), as well as cycle time (performance), measured across four collaborative trials and five time points, making group comparisons and the investigation of construct dynamics possible.
Results: Both control schemes demonstrated improved collaboration experiences compared to the baseline condition. Participants in the adaptive condition reported higher flow experience and workplace fit, and showed the fastest production times across all trials, while participants in the adaptable condition reported higher autonomy and task demands. Additionally, trust in the robot increases over time, harmonizing trust levels across conditions.
Discussion: Despite limitations related to an exploratory design and a small sample size potentially masking existing effects, the findings indicate that dynamic working conditions can improve worker experience. However, the findings do not present a consistent picture favoring either adaptive or adaptable control schemes. Further research is needed to determine which control scheme is preferable in different task contexts. To inspire future work, we derive a set of hypotheses from the study’s initial findings.

Citation:
Berretta, S., Tausch, A., & Glogowski, P. (2025). Maybe adaptive (not adaptable) automation in production: an experimental study comparing the locus of authority in work system dynamics. Frontiers in Organizational Psychology, 3, Article 1685961.

Station 3: Focusing on people in the changing landscape of occupations within professional fields

Jobs change when AI supports workers in their tasks. Whilst this can make tasks easier, it does not automatically mean that working conditions improve for employees. A human-centred approach ensures that work remains varied, offers opportunities for development, and enables people to acquire new skills. Employees must be able to develop their roles further when working alongside AI. The outcomes of work carried out for other people – for example, in healthcare or the service sector – must also become more reliable and of a higher standard through the use of AI.

Authors:
Uta Wilkens, Valentin Langholf, Marc Dewey

Abstract:
In this paper we analyse the role development of professionals in healthcare in face of AI applications to their workplaces. The conceptual background is role development theory aligned to human-AI work settings. The empirical fundament is a case study analysis conducted at Charité including a profile analysis of survey data from radiology (N=128) and a structured content analysis of ten semi-structured interviews with professionals. The outcome is the distinction of two most typical human-AI role concepts, (1) the AI-embracing human-AI role concept, and (2) the AI-ambivalent human AI-role concept. These types are based on the same set of antecedents in terms of AI literacy, former digital experience, individual perspective on the technology and the impact of AI on the overall change of individual tasks. This allows to understand why the first type experiences benefits from the human-AI role development while the second type cannot exclude personal harms. The AI-embracing role concept enhances role making with AI and incorporates AI implementation, the latent risk of AI in the AI-ambivalent concept leads to role taking against the technology.

Citation:
Wilkens, U., Langholf, V., & Dewey, M. (2024). Types of Human-AI Role Development – Benefits, Harms and Risks of AI-Based Assistance from the Perspective of Professionals in Radiology. Journal of Competences, Strategy & Management, 12.

Authors:
Sophie Berretta, Alina Tausch, Annette Kluge

Citation:
Berretta, S., Tausch, A., & Kluge, A. (2023). CollaborAId SMART: A concept for designing identity-creating basic work in the context of AI. In Gesellschaft für Arbeitswissenschaft e.V., Sankt Augustin (Ed.), Menschengerechte Arbeitsgestaltung – Basisarbeit und neue Arbeitsformen. GfA-Press.

Authors:
Valentin Langholf, Uta Wilkens

Abstract:
The use of artificial intelligence (AI) in work processes requires proactive changes to job roles, as areas of responsibility within job profiles are shifting, resulting in new patterns of interaction between humans and AI, and between employees. To avoid role conflicts, resistance to AI and other undesirable side effects of AI integration, organisations must support the development of human–AI roles through appropriate measures. This article presents a methodologically sound approach to role development (‘clarifying AI-augmented individual roles’ – clAIr), which was developed and tested using the example of service technicians in a mechanical engineering company, both before and during the introduction of AI-based services. It illustrates how role clarity in interaction with AI can be achieved when tasks are restructured, and how role development also encompasses internal collaboration with other departments and goal-oriented external communication within the customer environment. The method yields six basic roles, which are based on insights from role theory and encompass not only role clarity but also role identity and role innovation. clAIr enables a proactive, process-oriented examination of human–AI work roles.
Practical relevance: AI applications are affecting an ever-increasing number of work processes. Scientifically sound findings and examples of good practice are needed to ensure their successful integration into work processes. A socio-technical systems analysis focused on changes in work roles is particularly promising in this regard, as it translates the changes to tasks and occupations brought about by the use of AI into a comprehensive approach that also takes the identity of the worker into account. Existing job evaluation approaches focus on the characteristics of job designs, but are unable to capture the preceding process of role definition as a key determinant of implementation. This process support is enabled by the clAIr method for defining roles for working with AI. Its use requires an understanding of role theory and expertise in organisational development.

Citation:
Langholf, V., & Wilkens, U. (2024). Pathway to work with AI: Testing the clAIr role development method in an industrial work environment. Zeitschrift für Arbeitswissenschaft, 78, 377–386.

Station 4: The importance of job identity

Every job consists of a variety of tasks, which may be physical, social or technical in nature. These characteristics are relevant to our identity in the workplace, as they shape our sense of who we are, what matters to us and what makes us proud. When tasks are altered, shifted or replaced through the use of AI, this has an impact on our job identity. The transformation of work must therefore be shaped in such a way that identity-forming characteristics are preserved or enhanced.

Authors:
Sophie Berretta, Alina Tausch, Corinna Peifer, Annette Kluge

Abstract:
The use of artificial intelligence (AI) in the workplace makes it possible to organise work in such a way that it is carried out by human-AI teams. In order to evaluate efforts to design human-AI teaming workplaces, the JOPI (Job Perception Inventory) assessment tool is being developed as part of the ‘humAIne’ project. This inventory is intended to measure the effects of collaboration with AI on well-being, motivation and professional identity. From this, conclusions will be drawn regarding the successful and human-centred implementation of AI in the workplace.

Citation:
Berretta, S.; Tausch, A.; Peifer, C.; Kluge, A. (2022): Messung von Wohlbefindens-, Motivations- und Identitätsförderlichkeit von Mensch-KI-Teaming- Arbeitsplätzen. Magdeburg (online): GfA-Press (Technologie und Bildung in hybriden Arbeitswelten, 39).

Authors:
Sophie Berretta, Alina Tausch, Corinna Peifer, Annette Kluge

Abstract:
When work changes as a result of technological advances, this creates both challenges and opportunities for employees. To ensure that the use of artificial intelligence (AI) does not lead to the loss of the elements that give a job its identity, but rather maintains and promotes motivation and vigilance, the HUMAINE project is conducting research into human-AI collaboration. The Job Perception Inventory (JOPI) helps to ensure that the use of AI is designed with a human-centred approach.

Citation:
Berretta, S.; Tausch, A.; Peifer, C.; Kluge, A. (2022): Humanzentrierte Arbeitsgestaltung im Zeitalter von KI. ASU.

Authors:
Sophie Berretta, Alina Tausch, Corinna Peifer, Annette Kluge

Abstract:
Introduction: Artificial intelligence (AI) is seen as a driver of change, especially in the context of business, due to its progressive development and increasing connectivity in operational practice. Although it changes businesses and organizations vastly, the impact of AI implementation on human workers with their needs, skills, and job identity is less considered in the development and implementation process. Focusing on humans, however, enables unlocking synergies as well as desirable individual and organizational outcomes.
Methods: The objective of the present study is (a) to develop a survey-based inventory from the literature on work research and b) a first validation with employees encountering an AI application. The Job Perception Inventory (JOPI) functions as a work-analytical tool to support the human-centered implementation and application of intelligent technologies. It is composed of established and self-developed scales, measuring four sections of work characteristics, job identity, perception of the workplace, and the evaluation of the introduced AI.
Results: Overall, the results from the first study from a series of studies presented in this article indicate a coherent survey inventory with reliable scales that can now be used for AI implementation projects.
Discussion: Finally, the need and relevance of the JOPI are discussed against the background of the manufacturing industry.

Citation:
Berretta, S., Tausch, A., Peifer, C., & Kluge, A. (2023). The Job Perception Inventory: considering human factors and needs in the design of human–AI work. Frontiers in Psychology, 14, 1128945.

Station 5: Skills shortage – AI safeguards knowledge

The world of work is increasingly facing a shortage of skilled workers and the practical knowledge of the past. One example of this is the maintenance of old pumps and systems used to manage the long-term damage caused by mining. This requires specialist expertise, which is not always available. AI can help to bridge gaps in human knowledge and ensure pump performance. To achieve this reliably, it must draw on the vast knowledge archives of both the present and the past. This may also include elements such as water.

Authors:
Uta Wilkens, Julian Polte, Philipp Lelidis, Eckart Uhlmann

Abstract:
The paper specifies the genAI support needs for industrial maintenance against the background of a sociotechnical systems perspective. Emphasizing two needs, accessing implicit operator knowledge and prioritizing complex regulatory knowledge, a multi-layer architecture is outlined for an AI-based context-sensitive maintenance assistance system (MAS). The main purpose is to bridge knowledge gaps with genAI if human expertise and human implicit knowledge are not available and to cope with sub-process-specific challenges of multiple regulations. The MAS facilitates access to technical knowledge, distributes expertise, and shares implicit knowledge of experienced operators across different layers of information processing. The approach goes beyond standardization and has a high potential to enhance organizational as well as individual resilience.

Citation:
Wilkens, U., Polte., J., Lelidis, P., & Uhlmann, E. (2025). Bridging Knowledge Gaps with GenAI in Industrial Maintenance. Specific Needs and Contextualized Solutions. Industrie4.0 Science. 3/2025, S. 52-57. DOI: 10.30844/I4SD.25.5.XX

Station 6: Changes in the regional employment structure

New technologies give rise to new areas of employment, but they also lead to the loss of others. In the Ruhr region, this has led to structural changes on several occasions. Today, the Ruhr region is benefiting from the fact that new sectors, such as the healthcare industry and information technology, have become established and are ensuring stable employment growth.

Authors:
Monika Matzner, Linda Tuckwell, Alica Wilkens

Abstract:
The report presents the trend in employment in the Ruhr region from the era of rapid industrialisation to the present day in figures. It focuses on shifts in employment across economic sectors, illustrating these changes at selected points in time. The report also examines the interdependencies between economic sectors. To illustrate the changing importance of individual sectors for employment in the region, the figures are presented in relation to total employment.
This highlights the historical significance of mining for regional employment. In 1882, approximately 43 per cent of the region’s workforce was employed in mining. The significant importance of mining for regional employment was further reinforced by synergies with the iron and steel industry and the chemical industry. A strong industrial network emerged, attracting further workers who also contributed to the region’s growth.
The growth trend initially continued after the Second World War, despite a number of challenges. In response to the decline in employment in the mining sector, which began in the late 1950s, new industries such as the automotive sector were established. Overall, employment in mining and related industries still accounted for around 21 per cent by 1970. With the international oil crises, the shift into a difficult phase could no longer be held back. In the 1980s, there was a long-term decline in the industrial sector, coupled with job cuts, which led to a temporary rise in unemployment to around 15 per cent.
Over the last ten years, there has been another turnaround, this time for the better. Employment in the Ruhr region has risen by around 16 per cent. Healthcare professions, education and training, as well as IT and other ICT professions, have played a particularly significant role in this positive trend. This also includes employment related to the development and use of artificial intelligence (AI). New synergies are becoming apparent, for example between the healthcare sector and ICT professions. It is crucial that the region does not once again become overly reliant on a single key sector, as was once the case with mining. There are several growth sectors offering well-paid jobs and employing skilled workers. These sectors are developing both in tandem and independently of one another. None of the sectors accounts for more than 10 per cent of total employment.

Citation:
Matzner, M., Tuckwell, L., & Wilkens, A. (2025). Entwicklung der Beschäftigungsstruktur im Ruhrgebiet – Ausgewählte Etappen des Strukturwandels; Bericht im Rahmen der BMFTR-Projektförderung Kompetenzzentrum HUMAINE.

Station 7: Ethics and public participation in the use of AI

The use of AI brings with it risks and questions of ethical responsibility. Works councils and trade unions represent the interests of employees and are committed to ensuring that risks to employees are mitigated and that all parties benefit from technological change. This role is enshrined in law. The introduction of AI is subject to co-determination by works councils. Particularly in the Ruhr region, these councils can draw on past experience in managing structural change and advocate for a ‘transformation partnership’. The EU AI Act also ensures that AI is used in an ethically responsible manner.

Authors:
Thomas Haipeter, Manfred Wannöffel, Jan-Torge Daus, Sandra Schaffarczik

Abstract:
This article examines the role of employee participation in AI implementation, focusing on a case study from the German telecommunications sector. Theoretical discussions highlight concepts of employee participation and workplace democracy, emphasizing the normative basis for human-centered AI in Europe. The empirical analysis of the case study demonstrates social practices of human-centered AI and the importance of employee representatives and labor policies in sustainable technology. The contribution is structured into two main parts: first, discussing sociological concepts of employee participation and summarizing the role of works councils in shaping digital technology implementation. Second, focusing on a case study of AI regulations at Deutsche Telekom, highlighting the significant effects of employee participation and co-determination by the group works council in promoting socially sustainable AI implementation which is done via qualitative case analysis. The article highlights the significance of participation and negotiations and gives an example for social partnership relations in AI implementations.

Citation:
Haipeter, T., Wannöffel, M., Daus, J.-T. & Schaffarczik, S. (2024). Human-centered AI through employee participation. Frontiers in Artificial Intelligence, 7, 1272102.

Authors:
Alexander Ranft, Fabian Hoose, Claudia Niewerth, Matthias Preuß, Manfred Wannöffel

Abstract:
The introduction of artificial intelligence (AI) systems in workplaces poses new challenges for regulation and co-determination. Under the EU AI Act, binding requirements will come into force from 2025, which must be integrated at national level with the Works Constitution Act (BetrVG). The regional centre of excellence HUMAINE has developed a model AI works agreement (MBV KI) in accordance with Section 77 of the BetrVG, which strengthens works council rights and implements European regulatory practice in a practical manner. Supported by co-determination dialogues, the MBV KI allows for company-specific adaptation to ensure the responsible and human-centred use of AI. Using selected sections of the MBV KI as examples, this article demonstrates how a framework works agreement on AI can be structured in concrete terms and discusses its applicability to workplaces without a works council. The MBV KI presented here contributes to the sustainable shaping of the digital transformation, underpinned by social partnership.

Citation:
Ranft, A., Hoose, F., Niewerth, C., Preuß, M. & Wannöffel, M. (2026). Regulierung von humanzentrierter KI in Betrieben – Die HUMAINE Muster-Betriebsvereinbarung. Industry 4.0 Science, 42(1), S. 14–21.

Authors:
Manfred Wannöffel, Fabian Hoose, Alexander Ranft, Claudia Niewerth, Dirk Stüter

Abstract:
As part of the ‘humAIne’ regional centre of excellence, funded by the Federal Ministry of Research, Technology and Space (BMFTR), a process was developed using the ‘co-determination dialogues’ approach. This process enables management, employees and employee representatives to gradually build a shared understanding, through dialogue, of the complex challenges involved in the introduction of artificial intelligence (AI). Experiences from project partner companies – such as Doncaster’s Precision Castings-Bochum GmbH (DPC) – demonstrate how these co-determination dialogues not only help to develop legally binding regulations for a manageable, operationally embedded and sustainable practice of co-determination in the field of AI, but also initiate ongoing training processes for all stakeholder groups involved, in line with Articles 4 and 5 of the EU AI Act.

Citation:
Wannöffel, M., Hoose, F., Ranft, A., Niewerth, C. & Stüter, D. (2026). Mitbestimmungsdialoge zur humanzentrierten KI-Einführung – Dialogisches Verfahren der Entwicklung einer betriebsspezifischen Mitbestimmungspraxis. Industry 4.0 Science, 42(1), S. 92–98.

Station 8: HUMAINE – People with responsibility for AI

The HUMAINE team has researched and tested how to develop and use AI responsibly. The team includes members from academia, the professional world and regional development. Everyone shares responsibility for the ethical use of AI – including you. You can find out here what matters most to each team member.

What matters most to you?