Human Factor Analysis in the ALCHIMIA Project

The ALCHIMIA project is starting with the elicitation of the requirements of stakeholders, including shop-floor workers, to provide a solution that will be human-centred. The ambition of ALCHIMIA is to empower and augment workers’ capabilities through symbiotic robotic collaboration. A human-centric philosophy is guiding the design of the ALCHIMIA system, ensuring that workers’ capacities and skills are augmented by the envisioned Artificial Intelligence (AI) system, instead of trying to achieve a completely automated production process. Although AI models can automate several tasks, some concerns have been raised during the last years due to the uncertainty that their usage can cause in critical domains like manufacturing. AI applications and Machine Learning (ML) can result in opaque processes and outputs from the point of view of workers. Due to the great impact of the predictions on decision-making processes (e.g., in manufacturing), it is important to develop mechanisms and techniques that will provide necessary information to the users about the reasoning process of an AI model.

Cardiff University social scientists are leading the work packages which are aiming to understand organisational and employment implications (e.g., work organisation, worker autonomy, decision-making, tacit knowledge use, skill use, safety) of the ALCHIMIA system, including emerging skills and training needs, which will serve as key requirements for the whole project. Potential workplace issues that could arise from the ALCHIMIA system include increased surveillance/dataveillance, new skills and training needs, loss of expertise, changes to work organisation and job displacement.

Our ‘human factor’ analysis approach is informed by the European Sociological Association Ethics Guidelines for researching worker perspectives. We are adopting a phased approach with an ex-ante and ex-post framework. The first phase (ex-ante assessment) includes (recently completed) desk research to scope the field and then utilises surveys and qualitative interviews (individual and group) to understand current perspectives on the technology at each case site (which are currently being developed). The second phase (ex-post assessment) will build on phase one to deliver a post-insertion survey and evaluation interviews. The logic of the approach is to evaluate the prerequisites for the introduction of new technologies aimed at optimising energy use and minimising waste and the subsequent impact on workplace attitudes and practices. These methods are being used to understand how different stakeholders (i.e., operators, associated team members, supervisors and line managers, as well as trade unions and works council representatives) perceive and frame current workplace issues in relation to the introduction of new technological devices.

All aspects will be quantitatively and qualitatively assessed and summarised in a theoretical model of green technological effects and acceptance. This will inform the final recommendations for the design and implementation of the technology and training products, based on good practice templates. The results will ensure that the technology inserted is sensitive to all relevant social aspects to facilitate the development and application of the innovation in ‘human centric’ ways. Our research will help to develop guidelines for trust, safety and human use of AI tools in heavy industrial environments (including recommendations for human-centric technology development and insertion), a skills development strategy, and a training and education action plan.