Amid rapid advances in technological developments in interpreting, questions about the uniquely human dimensions of interpreting have gained renewed relevance (AIIC, 2025; Defrancq et al., 2024; Horváth, 2022; O'Hagan, 2019; Quintanero Raposo, 2024). While machine interpreting continues to advance, scholars argue that human interpreters still possess irreplaceable competences rooted in reflection, intuition, and emotional awareness during decision-making (Braun, 2019; Defrancq et al., 2024; Fünfer, 2014; Horváth, 2022). Building on this debate, this paper argues that simultaneous interpreting fundamentally relies on Emotional Labour (the regulation, negotiation, and performance of emotions in socially situated interaction (Hochschild, 2012/1979)), which remains difficult to model algorithmically.

The paper combines Hochschild (2012/1979)’s concept of Emotional Labour with sociological role theory (Dahrendorf, 2006; Schulz-Schäffer, 2024; Zwischenberger, 2013, 2017) to examine how interpreters make and justify decisions in line with role expectations. Empirically, the paper draws on cued retrospective protocols from six advanced MA students in conference interpreting. Using qualitative content analysis (Mayring, 2022), it explores how participants describe the reasoning behind their interpreting choices and how they negotiate the often competing expectations of speakers, audiences, and professional norms.

The analysis shows that interpreters do considerably more than transfer linguistic meaning. Their decision-making is characterised by continuous anticipation of audience needs, alignment with speakers' communicative intentions and emotional stance, and conscious regulation of their own emotional impulses. Emotional Labour appears not as an incidental aspect of interpreting but as a central mechanism through which interpreters exercise professional agency.

Rather than asking whether AI will replace interpreters, the paper shifts the discussion towards identifying the dimensions of interpreting that resist algorithmic modelling. It proposes Emotional Labour as a conceptual framework for understanding the distinctly human value of interpreting and argues that this perspective can enrich current debates on human agency, AI, and professional identity.