Process industry disrupted: AI and the need for human orchestration

According to EU policy makers, the introduction of AI within Process Industry will help big manufacturing companies to become more sustainable. At the same time, concerns arise about future work in these industries. As the EU also wants to actively pursue human-centered AI, this raises the question how to implement AI within Process Industry in a way that is sustainable and takes views and interests of workers in this sector into account. To provide an answer, we conducted ‘ethics parallel research’ which involves empirical research. We conducted an ethnographic study of AI development within process industry and specifically looked into the innovation process in two manufacturing plants. We showed subtle but important differences that come with the respective job related duties. While engineers continuously alter the plant as being a technical system; operators hold a rather symbiotic relationship with the production process on site. Building on the framework of different mechanisms of techno-moral change we highlight three ways in which workers might be morally impacted by AI. 1. Decisional - alongside the developmental of data analytic tools respective roles and duties are being decided; 2. Relational - Data analytic tools might exacerbate a power imbalance where engineers may re-script the work of operators; 3. Perceptual - Data analytic technologies mediate perceptions thus changing the relationship operators have to the production process. While in Industry 4.0 the problem is framed in terms of ‘suboptimal use’, in Industry 5.0 the problem should be thought of as ‘suboptimal development’.