Competing views of generative AI in education — as transformational pedagogy, productivity tool, or surveillance risk — are driving divergent institutional responses that will determine whether AI augments teachers, reshapes costs, or widens inequalities; policymakers should prioritize teacher upskilling, public goods, and flexible governance to steer outcomes.

Main Finding

Using the Social Construction of Technology (SCOT) framework to review literature from 2021–2025, the study finds that divergent stakeholder interpretations of generative AI (GenAI) in education—structured around three core interpretive tensions—are actively shaping how GenAI is developed, governed, and adopted in educational settings. These competing views produce distinct institutional responses (AI literacy, human-led learning models, governance frameworks) and a dynamic conceptual model explaining the social construction of GenAI in education.

Key Points

• Three dominant interpretive tensions drive debates about GenAI in education:
  • Transformational force vs. transactional tool: Is GenAI a paradigm-shifting pedagogical technology or primarily a productivity/efficiency aid?
  • Cognitive augmentation vs. cognitive atrophy: Does GenAI enhance cognitive skills and higher-order thinking or erode learning and critical skills through over-reliance?
  • Digital empowerment vs. digital enslavement: Does GenAI expand access and learner agency or increase surveillance, dependence, and inequality?
• Stakeholders examined include students, teachers, administrators, policymakers, edtech firms, and parents — each with distinct technological frames and priorities.
• Interpretive flexibility is high: the same GenAI capabilities are framed differently by groups, generating controversies over assessment, authorship, academic integrity, pedagogy, and labor roles.
• Institutional responses emerging in the literature:
  • AI literacy and critical digital skills programs to enable informed use;
  • Human-led learning models that preserve teacher agency and emphasize human–AI complementarity;
  • Governance frameworks focusing on responsible deployment, transparency, and accountability mechanisms.
• The paper proposes a conceptual model mapping how stakeholder interpretations, controversies, and institutional responses co-evolve to shape GenAI adoption trajectories in education.

Data & Methods

• Method: Interpretive literature review synthesizing peer-reviewed and relevant grey literature published 2021–2025.
• Theoretical lens: Social Construction of Technology (SCOT), using core constructs:
  • Relevant social groups: actors who ascribe meaning and interests to the technology (students, teachers, etc.).
  • Interpretive flexibility: the multiplicity of meanings ascribed to GenAI features.
  • Technological frames: stakeholders’ problem definitions, goals, and norms relating to GenAI.
  • Controversies: focal points of dispute that direct development and regulation.
• Analytical approach: The review codes and interprets recurring themes and tensions across studies rather than aggregating quantitative effect sizes.
• Limitations (inherent to the method): dependent on existing literature coverage and framing; interpretive synthesis cannot establish causal estimates or precise economic magnitudes.

Implications for AI Economics

• Human capital and labor complementarities
  • GenAI shifts the skill mix demanded in education: greater returns to higher-order cognitive skills, AI literacy, and pedagogical design, while routine grading or content generation tasks may be automated.
  • Economic models should treat teachers and GenAI as potential complements (augmenting teacher productivity) or substitutes (displacing tasks), with outcomes depending on institutional choices driven by stakeholder frames.
• Productivity and cost structure of education
  • Adoption trajectories influenced by interpretive tensions will determine whether GenAI lowers marginal costs (through automation) or mainly enhances quality (raising value but not reducing costs).
  • Investment in AI literacy and governance creates upfront fixed costs; the distribution of these investments affects long-run average costs across providers and countries.
• Market structure and competition
  • Divergent stakeholder-driven pathways (e.g., dominant edtech platforms vs. institution-led deployments) could lead to market concentration around major GenAI suppliers or a more decentralized ecosystem; regulatory responses will affect entry barriers and pricing power.
• Distributional effects and inequality
  • If adoption favors resource-rich institutions that can implement human-led models and robust governance, GenAI risks widening educational and labor-market inequality.
  • Conversely, well-designed public interventions (subsidies for AI literacy, open-source models, or shared governance standards) could democratize benefits.
• Incentives, governance, and externalities
  • Misaligned incentives (publish/assessment pressures, edtech firms’ commercial motives) can produce negative externalities: credential inflation, cheating, surveillance-driven behavior, or skill atrophy.
  • Economic policy instruments (standards, accreditation rules, procurement criteria, conditional subsidies) can internalize externalities and steer adoption toward socially desirable equilibria.
• Measurement and empirical research priorities for AI economics
  • Need for causal estimates of GenAI’s effect on learning outcomes, teacher productivity, and labor-market returns using experiments, quasi-experiments, and longitudinal data.
  • Cost–benefit and distributional analyses that account for governance costs, training, and adaptation timescales.
  • Market analyses of competition, pricing, and platform dynamics in the GenAI-enabled edtech sector.
  • Macro-level modeling of how skill-biasing technological change in education interacts with labor demand and inequality.
• Policy design guidance
  • Prioritize investments that enhance complementarities (teacher upskilling, curriculum redesign) rather than purely automating tasks.
  • Support public goods (AI literacy curricula, benchmarks, oversight capacity) to mitigate unequal access and negative externalities.
  • Design regulatory frameworks that are flexible to interpretive diversity but enforce basic accountability (transparency, data protection, assessment integrity).

Summary takeaway: The social construction of GenAI in education—shaped by competing stakeholder frames and institutional responses—will critically determine its economic impacts on productivity, labor demand, inequality, and market structure. AI-economics research and policy should therefore focus on measuring complementarity vs. substitution effects, institutional costs of governance and literacy, and distributional outcomes to inform interventions that capture benefits while limiting harms.