“It’s all just code”: How is responsibility embedded (or not) in the work of a team of software engineers?

Recent research has documented how software can sometimes contribute to harm and inequality, affecting individuals, societies, and the environment. We argue that human values embedded in software, consciously or unconsciously during design and development, play a significant role in producing these outcomes. In other words, to understand responsibility in software engineering, it is crucial to understand how Software Engineers (SEs) incorporate responsibility into their practices, and if or how they reflect on the potential impact of their work. The research reported here probes this assertion by focusing on a team of research SEs involved in developing a commercial source code clone detection tool. Through reflexive thematic analysis of in-depth interviews, with meta-reflections from the SEs Principal Investigator, we offer a detailed account of the practices and artefacts referenced when these professionals reflect on responsibility in their work. Our analysis suggests that the team does not consider responsibility as a singular, unified concept. Instead, reflections on potential negative software impacts and mitigation strategies vary depending on whether they relate to the institutional/organisational context, the specific task, or the intended use. Overall, these SEs tend to perceive software as neutral or inherently beneficial, which, in turn, feeds into a lack of practice of engaging in speculation about the broader implications of the technologies they work with. To address these gaps, we propose a paradigm shift toward an infrastructural view of responsibility - interdisciplinary, sociotechnical, and situated – embedded within the diverse practices and organisational cultures of SE teams operating in different contexts.