The Social Penalties and Rewards of AI-Augmented Work.

While generative AI (GenAI) is often framed as a tool for individual productivity, its integration into the workplace can become a social disruption. This study focuses on the moral economy of AI-augmented labour, proposing that GenAI use is a normatively ambiguous act that can trigger distinct social penalties and rewards within organisations. By analysing how workers evaluate the AI-augmented workflows of others, varied across multiple dimensions, this study empirically investigates whether we are witnessing the development of a new workplace social order.
Building on prior research into workplace deviance, cyberloafing, and slack attribution theory, we develop an experimental factorial design to explain how employees are socially rewarded or penalised for using GenAI. The core theoretical claim to be tested is that evaluations around GenAI use are driven by a configuration of different signals. These signals are cognitively integrated through attribution processes and translated into social and organisational sanctions or rewards.
Empirically, this argument is tested using five experimental vignette studies with a total of approximately 2000 currently employed UK respondents. The designs implement a balanced between-subject factorial structure, manipulating different combinations of dimensions: Output quality, legitimacy, transparency, task impact, impact level, time use, loafing, remote work, age, gender, tenure and qualification. The focus of the analysis lies on main effects with interactions applied where plausible.
These factors are conceptualised as possibly shaping attribution mechanisms, including perceptions of competence, effort, honesty and orientation towards collective versus individual interests. Dependent variables capture normative judgements of the acceptability of GenAI use, social evaluations (team fit and helping intentions), and preferences for organisational penalties and rewards (recognition and workload adjustments). Additional models also focus on individual level predictors, including AI attitudes, work norms and demographic characteristics of the respondent.
By explicitly modelling how multiple signals are integrated into attribution processes, this study contributes to sociological theories of evaluation and deviance under conditions of technological change. It demonstrates how individually beneficial uses of GenAI may generate collective tensions. More broadly, the findings highlight how different factors might shape the social legitimacy of emerging technologies, with implications for inequality, coordination and the governance of AI enabled work.