ALCHIMIA Project Presentation at the Artificial Intelligence and Transformations of Work Conference 

By Rachel Hale, Dean Stroud, Martin Weinel and Vinicio Di Iorio

The ALCHIMIA Project was presented at the Artificial Intelligence and Transformations of Work Conference in Grenoble, France, in November.  The aim of the conference was ‘to shed light on issues related to the transformation of human work, providing a contribution to these issues from an ethical perspective. To do this, it is necessary to revisit the place of work for the human condition and analyze the redefinitions underway, asking: is AI redefining work and its value?’.  The following disciplines were invited to participate in the reflection: philosophy, sociology, social history, economics, information and communication sciences, social psychology, ergonomics, computer science, robotics.  The conference was structured around four axes: 1) Replacement: myth or reality; 2) Case Studies: analyzing the role of current systems; 3) Fairness and justice at work: what values for an ethics of work?; and, 4) Conditions of Work.

The ALCHIMIA Project was included in one of the case study sessions.  Our presentation was on the role of social scientists in developing the ALCHIMIA artificial intelligence platform for steelmaking work. We considered how, as sociologists and STS (Science and Technology Studies) scholars, our presence in this interdisciplinary project is allowing us to examine how AI as complex sociotechnical systems are situated within the practices that they emerge from, including our practices as social scientists on the project.  The presentation outlined not only our formal role in the project (which we have written blogs on before), but also how as social scientists we are taking a critical empirical sociology/STS approach to specific ‘Big Data’ practices, AI and waste work, related to scrap steel.

We talked about our experience in the ALCHIMIA project of the opportunities and challenges of doing interdisciplinary research, doing research in large consortia and doing social science fieldwork in industrial settings.    The photograph below was taken during the fieldwork visit at Celsa in Poland in May, while we were conducting an interview in the working scrapyard with the scrapyard manager. Here, we observed trucks delivering and unloading scrap metal and then large magnets moving the metal to the appropriate silos, after the scrap steel had been classified by the scrap classifier. The photograph, and the requirement for us to wear safety gear, illustrates some of the issues we encountered in doing fieldwork in industrial settings, such as conducting and recording interviews in noisy, polluted, and hazardous settings.  Some of the interviews took place in the control rooms of the foundries, which were noisy, hot and the rooms vibrated because they were situated above and to the side of the crucibles.  In these interviews, some of the participants were simultaneously doing their jobs as well as answering interview questions, which allowed us to observe steelmaking processes in situ.  This was essential for our understanding of the social requirements of the AI that we are involved in developing, including the circumstances under which the input and output data is generated that the AI is being trained on.