题目:AI Artifact for Managers: How Generative AI Complements or Substitutes for Human Leadership
主讲人:许博 副教授
复旦大学管理学院信息管理与商业智能系
时间:2026-11-24 13:30-14:30
地点:线下 - 史带楼410室
线上 - 腾讯会议号 439 595 539 / 密码722722
内容摘要:
Information systems research has largely examined how employees and customers respond to AI agents, leaving AI used by managers largely unexamined. This paper scrutinizes whether AI affect managers’ job effectiveness and how generative AI may substitute transactional leadership skills (the ability to generate high-quality feedback) but complement transformational leadership skills (the ability to motivate employees to accept the feedback advice). We leverage a randomized field experiment with 369 employees and 12 managers in a fintech call center, where a generative AI system identified errors and produced corrective feedback that managers delivered verbatim to coach employees without disclosing the AI’s involvement. We find that although before AI adoption, transactional managers are more effective than transformational managers in engendering superior employee performance, after AI adoption, AI substitutes for transactional leadership, and transformational managers become more effective. This reversal finding is robust to wild cluster bootstrap, exact randomization inference, perfect manager-level rank separation, leave-one-manager-out estimation, and manager fixed effects. Mechanism analyses support that without AI, transactional leadership produces higher-quality feedback content leading to higher effectiveness than transformational leadership; but after AI adoption, we find not only that such feedback quality gap is eliminated (explaining the substitution between AI and transactional leadership), but also that employees are more likely to be motivated to seek out and adopt the feedback delivered by transformational managers (accounting for the complement between AI and transformational leadership). Additional analyses find that the complement is amplified among higher-prior-performing employees and spills over into higher organizational commitment and lower turnover intentions. We contribute to IS research by theorizing and revealing causal evidence of a new AI artifact for value creation in managerial work wherein AI reshapes the returns to human leadership in organizations.

复旦大学管理学院信息管理与商业智能系