Generative AI can act as a socio-digital peer in teacher education, not just a tool, by providing motivational, emotional, and dialogic support that reshapes how pedagogical skills are formed and how institutions adopt and capture value.

Main Finding

Generative AI in higher education functions beyond an administrative or instructional tool: when integrated as a socio-digital peer learner it can reshape teacher education by establishing motivational climates, offering emotional/non‑judgmental support, and co-constructing pedagogical knowledge. The study argues for a paradigm shift from “technology-as-tool” to “technology-as-social-collaborator,” with implications for curriculum, teacher identity, and dialogic agency alongside non‑human peers.

Key Points

• Instrumentalist framing (administrative assistant or tutor) understates GenAI’s ontological potential as a social actor in collaborative learning.
• Using Constructivist Grounded Theory, the authors develop a substantive theory positioning GenAI as a socio-digital peer learner that:
  • Creates motivational climates and sustained engagement.
  • Provides emotional support and a non-judgmental space for practice and reflection.
  • Participates in co-construction of pedagogical knowledge (dialogic, iterative interaction).
  • Enables future educators to develop “dialogic agency” — the capacity to reason and negotiate pedagogical choices in tandem with non-human peers.
• Methodological triangulation includes repeated interviews with global pedagogical and technical experts, synthetic data from recursive LLM dialogues, and industry field observations.
• The resulting framework recommends curriculum transformation in teacher education to incorporate collaborative practice with GenAI agents rather than merely teaching tool-use.

Data & Methods

• Theoretical approach: Constructivist Grounded Theory (iterative coding, theory generation from data).
• Data sources:
  • Repeated semi-structured interviews with a heterogeneous, global sample of pedagogical and technical experts.
  • Synthetic/experimental data: recursive dialogues with advanced LLMs to simulate interaction dynamics and generate analytic material.
  • Industry-based field observations documenting real-world deployments and practices.
• Triangulation: cross-validation across human expert perspectives, model-based interactions, and field evidence to substantiate emergent categories (motivation, emotional support, non-judgmental zone, co-construction).
• Methodological strengths noted: iterative, theory-building design; novel use of LLM dialogues as part of empirical triangulation.
• Potential methodological limits (implicit): reliance on expert samples and synthetic dialogues may bias toward optimistic views of GenAI capabilities; study is substantive/theory-generating rather than causal or quantitative.

Implications for AI Economics

• Human capital formation and productivity
  • GenAI-as-peer could change the content and pace of teacher skill acquisition, potentially accelerating pedagogical skill formation and altering returns to in-service training.
  • If effective, these tools could raise measured educational “productivity” (faster skill acquisition per unit cost), affecting aggregate labor productivity estimates in education-related sectors.
• Labor demand and task reallocation
  • GenAI that performs social, motivational, and co-creative functions is more of a complement (augmenting teacher capabilities) than a pure substitute, but could still displace some tasks (e.g., routine feedback) and shift demand toward higher-order pedagogical competencies.
  • Demand may rise for roles that manage AI-human interaction, curriculum redesign, and ethical oversight.
• Market structure, value capture, and competition
  • Firms that provide socio-interactive GenAI platforms could capture significant value through proprietary models and content, creating platform dynamics and potential concentration in educational AI markets.
  • Open models or interoperable standards could mitigate concentration; policy choices will affect who captures gains from productivity improvements.
• Pricing, adoption, and diffusion
  • Adoption incentives depend on measurable learning outcomes, cost reductions, and liability/credentialing frameworks. Evidence of improved teacher readiness may spur institutional procurement.
  • Network effects: widespread integration in teacher education could standardize AI-mediated pedagogies, accelerating uptake in K–12 and higher education markets.
• Inequality and distributional effects
  • Potential to democratize access to high-quality practice environments, but unequal access to advanced models or institutional support may widen disparities across institutions, regions, and socioeconomic groups.
• Measurement and evaluation challenges
  • New metrics needed to evaluate dialogic agency, emotional support quality, and long-term pedagogical outcomes attributable to AI-peer interactions.
  • Economic evaluations should go beyond short-term test-score gains to include effects on teacher retention, classroom practices, and student long-run outcomes.
• Policy and governance implications
  • Regulation and standards for assessment, safety, and transparency will affect adoption costs and risk pricing.
  • Public investment in open or public-interest models could alter market dynamics and distribution of benefits.

Suggested research questions for AI economics arising from this study - To what extent do socio-digital peer systems substitute for versus complement different teacher tasks? How does that affect wages and employment composition? - What are the returns to investing in AI-mediated dialogic agency compared with traditional teacher training? - How does the introduction of GenAI peers change institutional adoption decisions, pricing strategies of vendors, and market concentration? - What are the welfare and distributional consequences of scaling socio-digital peers across diverse education systems?

Overall, the paper suggests a shift in how economists should model GenAI in education: from productivity tools with simple task-based substitution/complementarity relationships toward agents that reshape skill formation, institutions, and market structures through social and dialogic interactions.