AI in industrial work: perspectives and design options

In the new report ‘Artificial intelligence and industrial work - perspectives and design options’ by the Industry 4.0 Research Advisory Board, the Fraunhofer Institute for Mechatronics Design Technology IEM and the Fraunhofer Institute for Material Flow and Logistics IML examine the integration of artificial intelligence into industrial work.

The study is based on a comprehensive literature review as well as a total of 25 interviews and two workshops with experts, particularly from industry. The expertise focuses on the mechanical and plant engineering, vehicle construction and automation technology sectors with regard to the use of AI in the production environment. The interviews were divided into three subject areas: the use of AI-based systems, the consequences of the use of AI-based systems for employee deployment and the design approaches and development prospects of AI-supported industrial work. Focus group discussions were held in the two workshops in order to gain even deeper and more comprehensive insights based on the interview results [1].

AI as a future colleague of humans

Artificial intelligence has the potential to initiate a second wave of Industry 4.0. As with previous research on Industry 4.0, the question of the transformation of industrial work and its design options is therefore becoming increasingly important. Previous research has shown that the use of technology can be problematic if it leads to monotonous activities as a result of automation gaps. On the other hand, it is desirable if the remaining activities do not lead to either excessive or insufficient demands on people. Compared to the previous automation and digitalisation of industrial production, AI can carry out unpredictable and unplanned processing steps, making it an equal player to humans - a kind of colleague. This collaboration is also known as hybrid intelligence. Due to the different strengths of humans and AI, their complementary cooperation can bring considerable advantages [1], [3], [4], [5].

AI is in the starting blocks - quality checks are particularly widespread

According to the interviews conducted, the introduction of AI is in the starting blocks in many places. 44% of the specialist lights surveyed report that they are already familiar with an AI application in productive use. In most cases, this is a single or a few applications. A further 36% of the experts surveyed stated that AI is currently being introduced or piloted. The remaining 20 % are not yet using AI at all. This confirms the assumption that artificial intelligence is increasingly penetrating industrial production, but is not yet widespread. The specific fields of application of AI in the production environment are shown in Figure 1. AI applications are already being used particularly frequently in the area of quality control, for example in optical quality inspections using automated image analysis. Another frequently mentioned field of application is the monitoring of production processes, such as AI-supported predictive maintenance. Overall, the majority of experts expect AI applications to be in use in all eight fields of application listed in five years' time. This also applies to areas such as the planning and development of production systems, in which AI has been used rather little to date, such as AI-supported ergonomics analyses [1].

Changing job profiles through AI

The experts expect that AI will take over some activities in most jobs in the future, but that there will be no complete substitution of job profiles. Four specific job profiles were analysed in more detail with regard to the impact of AI: warehouse logistics specialist, PLC programmer, production planner and assembly assistant. Possible changes to the job profiles as a result of various AI applications were discussed and both a best-case and worst-case scenario were created for each job profile. The experts expect efficiency and productivity to increase in all job profiles through the use of AI.

For other aspects such as skills requirements or job satisfaction, however, the assessments are ambivalent. The study shows that the effects on employees cannot be determined solely by the AI applications introduced. Work organisation and accompanying measures play a decisive role, as does the specific design of the technology. Possible developments in the world of work are shown in Figure 2. Instead of a substitution or upgrading of job profiles, however, a polarisation of work is assumed: the job profiles of production planners and PLC programmers will tend to be upgraded, while the job profiles of warehouse logistics specialists and assembly assistants will tend to be downgraded [1], [2].

Fields of action for human-centred AI design

In order to realise a human-centric AI design and prevent the devaluation of job profiles, for example, the experts were asked about recommended measures. These were categorised into nine fields of action, which are shown in Figure 3 (next page) and provide a comprehensive basis for the successful and humane integration of AI into industrial work processes.

The programmes and measures within a company should be anchored in the form of an AI strategy, which in turn can be part of the digitalisation or corporate strategy. A clear organisation of the AI introduction is necessary to ensure both top-down coordination and the bottom-up collection of ideas. The co-design of the AI introduction by affected employees contributes to the creation of acceptance, but above all also ensures the utilisation of empirical knowledge. The availability and quality of data are fundamental to the use of AI, so companies must ensure that they have the necessary data.

Before introducing AI, existing processes should be analysed and evaluated in order to identify potential for optimisation through AI. A suitable collaboration and culture are necessary in order to pool data across departments and gain new insights through the involvement of different domains. Raising awareness and transparency about the goals and functionality of the AI introduction are important in order to avoid fears or false expectations. The development and promotion of internal competences are crucial to prepare employees for working with AI and to fully exploit its potential. Utilising external skills can help companies to close gaps in AI knowledge and better scale solutions [1].

Conclusion

Individual activities within job profiles can be almost completely automated by AI in the future. In all likelihood, however, this will neither lead to the elimination of jobs nor to a reduction in work activities, as new activities will be created by AI alongside those that will be eliminated. It is therefore to be expected that, above all, the structure of job profiles will change. These should be designed for equal interaction between humans and AI and contribute to satisfaction, personal development and social compatibility. In view of the rapid developments in the field of (generative) AI, it is crucial for companies to act proactively and create a working environment that fulfils technological as well as human and social requirements in order to secure their competitiveness. The expertise of the Research Advisory Council is available for free download on the acatech website.

Fields of application of AI in the production environment [1].

Development scenarios for AI-supported work [1], [2].

Fields of action based on the experts' assessments [1].

References

[1] Forschungsbeirat Industrie 4.0/acatech – Deutsche Akademie der Technikwissenschaften (Hrsg.). (2024). Künstliche Intelligenz und industrielle Arbeit – Perspektiven und Gestaltungsoptionen. DOI: 10.48669/fb40_2024-1

[2] Hirsch-Kreinsen, H., Ten Hompel, M., Ittermann, P., Falkenberg, J., & Dregger, J. (2016). Social Manufacturing and Logistics: Konturen eines Leitbildes digitaler Industriearbeit: Bericht des Forschungsprojektes“. Technische Universität Dortmund, Wirtschafts-und Sozialwissenschaftliche Fakultät, Forschungsgebiet Industrie-und Arbeitsforschung.

[3] Huchler, N. (2022). Komplementäre Arbeitsgestaltung. Grundrisse eines Konzepts zur Humanisierung der Arbeit mit KI. Zeitschrift für Arbeitswissenschaft, 76(2), 158-175.

[4] Jarrahi, M. H. (2018). Artificial intelligence and the future of work: Human-AI symbiosis in organizational decision making. Business horizons, 61(4), 577-586.

[5] Schlick, C., Bruder, R., & Luczak, H. (2018). Arbeitswissenschaft. Springer-Verlag.