Transformational leadership still links to better employee outcomes, but in AI-intensive workplaces leaders increasingly act as designers of human–AI systems; economists should model leadership as managerial capital that shapes AI adoption, productivity and distributional effects.

Main Finding

Transformational leadership—traditionally framed as a human interpersonal influence process—retains robust links to positive employee attitudes and performance but must be reconceptualized for AI-intensive, digital workplaces. The literature review shows (1) four decades of meta-analytic support for transformational leadership outcomes, mediators, and moderators, (2) substantive critiques of its conceptual and measurement foundations, and (3) an emergent shift toward framing leadership as the design and orchestration of human–AI socio-technical systems rather than only dyadic human influence.

Key Points

• Historical lineage: Theories from Burns and Bass established transformational leadership as a leader-driven process that inspires followers through vision, intellectual stimulation, individualized consideration, and charisma.
• Meta-analytic evidence: Across ~40 years of meta-analyses, transformational leadership is consistently associated with positive follower attitudes (e.g., commitment, job satisfaction), extra-role behaviors, and performance; effects operate through cognitive and affective mediators and vary with contextual moderators.
• Major critiques: Concerns include conceptual overlap with other leadership constructs, construct breadth and fuzziness, measurement dependence on the Multifactor Leadership Questionnaire (MLQ), and limited clarity on causal mechanisms (the "follower-transformation" black box).
• Digital-era reconceptualization: Emerging e-leadership and digital leadership literatures treat leadership as distributed, mediated by information technologies, and increasingly entangled with AI systems that can act as decision aids, communication intermediaries, or partial substitutes for leader tasks.
• Measurement gaps: Heavy reliance on the MLQ is increasingly insufficient in digital contexts; new measures should capture leader use of AI tools, distributed decision architectures, communication through digital channels, and human-AI interaction quality.
• Proposed research agenda (four themes): (a) theorize the follower-transformation mechanism, (b) reconceptualize leadership as design of human–AI symbiotic systems, (c) expand measurement beyond the MLQ, and (d) situate transformational leadership across cultural and technological contexts.

Data & Methods

• Study type: Systematized literature review synthesizing theoretical, meta-analytic, and critical literatures on transformational leadership and its digital/AI reconceptualizations.
• Evidence synthesized: Four decades of meta-analyses and primary theoretical/empirical studies on outcomes, mediators, moderators, and critiques; emergent empirical and conceptual work in e-leadership and digital leadership.
• Analytic approach: Conceptual lineage tracing, synthesis of aggregated meta-analytic findings (outcomes, effect sizes, mediators, boundary conditions), and critical analysis of measurement and validity concerns; identification of gaps and agenda-setting for future research.
• Methods recommended by authors (for future work): Mixed methods including field experiments, longitudinal designs, network analysis, digital trace data, and new psychometric tools tailored to human–AI work environments.

Implications for AI Economics

• Rethinking complementarities: Transformational leadership should be modeled as managerial capital that interacts with AI capital. Economists must account for leader-driven organizational choices (e.g., task allocation, AI deployment, incentives, training) when estimating AI productivity complementarities with labor.
• Endogeneity and causal inference: Leadership effects are endogenous to firm selection and internal processes. Studies on AI adoption and productivity should instrument or exploit quasi-experiments to separate leader-driven adoption/performance effects from other firm characteristics.
• Measurement and data opportunities: Digital workplaces generate rich trace data (communication logs, workflow metadata, AI system usage). Combining these with traditional leader surveys can better capture leadership behaviors in AI-mediated settings and enable micro-level causal analysis (e.g., leader messages, AI recommendations, follower responses).
• Policy and labor-market effects: Leadership that effectively designs human–AI work systems may mitigate displacement risks, reshape task content, and influence retraining returns. Policymakers and firms should consider leadership development as part of strategies to realize equitable gains from AI.
• New empirical questions for AI economists:
  • How do transformational leadership practices affect firm-level AI adoption rates, AI-driven productivity gains, and returns to skill?
  • What are the causal channels (e.g., trust-building, task reallocation, learning incentives) through which leaders influence human–AI complementarity?
  • Do leaders amplify or attenuate distributional effects of AI (wage inequality, employment volatility) across occupations, sectors, and cultural contexts?
  • Can AI systems partially substitute for routine elements of leadership (e.g., monitoring, information aggregation) and what are the productivity and welfare consequences?
• Methodological suggestions:
  • Use firm-panel and worker-panel datasets linked to leadership measures and AI adoption indicators to estimate dynamic complementarities and heterogeneous effects.
  • Leverage natural experiments (e.g., staggered AI deployments, policy changes, leadership turnover) and randomized field interventions (leadership training on AI integration) to identify causal impacts.
  • Exploit digital trace data and network methods to operationalize concepts like distributed leadership, communication patterns, and human–AI interaction quality.
  • Develop and validate new measurement instruments beyond the MLQ that capture leader behaviors specific to AI-mediated work (e.g., governance of algorithmic decision-making, fostering AI trust, designing human-AI workflows).

Takeaway for AI economists: Incorporate leadership as an active, measurable component of the production function when studying AI adoption and impacts; design empirical strategies that capture how leaders shape human–AI complementarities, distributional outcomes, and firm performance.