Human or Algorithmic Managers in Expert Knowledge Work? The Interplay of Management Style and Managerial Function

Algorithmic management is entering expert knowledge work, but when workers accept algorithmic rather than human managers remains unclear. We surveyed Finnish expert knowledge workers (N = 1,156) using a factorial scenario experiment varying manager type (human/algorithmic), management style (controlling/balanced /flexible), and six managerial functions and measured leader–member exchange (LMX), organizational trust, AI familiarity and work complexity. Findings cover five functions; rewarding was excluded due to insufficient model fit. Management style was the strongest predictor: flexible styles were rated highest and controlling lowest, regardless of manager type. Human managers were preferred for goal setting, performance evaluation, and supportive leadership; algorithms only for scheduling under controlling conditions. Monitoring showed no consistent difference. High-quality LMX buffered reactions to controlling styles, and in performance evaluation and supportive leadership its association with acceptability was stronger for algorithmic managers, narrowing the human advantage. Organizational trust and LMX were more consistently implicated than AI familiarity, qualifying algorithm-aversion accounts.