Agentivism: a learning theory for the age of artificial intelligence
Authors: Lixiang Yan, Dragan Gašević
Organizations: School of Education, Tsinghua University. · Faculty of Education and School of Computing & Data Science, The University of Hong Kong. · Faculty of Information Technology, Monash University.
Learning theories have historically changed when the conditions of learning evolved. Generative and agentic AI create a new condition by allowing learners to delegate explanation, writing, problem solving, and other cognitive work to systems that can generate, recommend, and sometimes act on the learner's behalf. This creates a fundamental challenge for learning theory: successful performance can no longer be assumed to indicate learning. Learners may complete tasks effectively with AI support while developing less understanding, weaker judgment, and limited transferable capability. We argue that this problem is not fully captured by existing learning theories. Behaviourism, cognitivism, constructivism, and connectivism remain important, but they do not directly explain when AI-assisted performance becomes durable human capability. We propose Agentivism, a learning theory for human-AI interaction. Agentivism defines learning as durable growth in human capability through selective delegation to AI, epistemic monitoring and verification of AI contributions, reconstructive internalization of AI-assisted outputs, and transfer under reduced support. The importance of Agentivism lies in explaining how learning remains possible when intelligent delegation is easy and human-AI interaction is becoming a persistent and expanding part of human learning.
Figures & tables
Theory
Can explain
Falls short
Behaviourism
Why AI use can become reinforcing through immediate feedback, fluency, speed, effort reduction, and repetition Watson (1913) ; Skinner (1938) ; Ngu et al. (2025) ; Pan et al. (2025) .
Whether repeated AI-supported success reflects durable learning or merely reinforced dependence; why slowing users down through justification can reduce over-reliance Bučinca et al. (2021) .
Cognitivism
How AI changes cognitive load, memory demands, schema construction, retrieval, and offloading during learning Miller (1956) ; Atkinson and Shiffrin (1968) ; Stadler et al. (2024) ; Bastani et al. (2025) ; Liu et al. (2025) .
When offloading supports learning versus substitutes for it; how AI-assisted performance becomes retained, transferable human competence.
Constructivism
Why dialogue, interpretation, participation, and meaning making remain central when learners interact with AI conversationally Dewey (1938) ; Piaget (1952) ; Vygotsky (1978) ; Kim et al. (2026) ; Jin et al. (2025) ; Xing et al. (2026) .
Whether AI-generated guidance is epistemically trustworthy; interaction alone does not guarantee justified learning Salvi et al. (2025) .
Connectivism
Why learning depends on distributed networks of people, tools, and information resources Siemens (2005) .
What changes when network nodes become generative, persuasive, and partially agentic, shaping representation, priorities, and action in real time Wang et al. (2023) ; Tan et al. (2026) .
Table 1: What classical learning theories can explain, and where they fall short in the age of generative and agentic AI
Figure 1: Conceptual model of Agentivism as a learning theory for human-AI interaction. Agentivism explains learning as durable growth in human capability under conditions where generative or agentic AI contributes directly to task completion. The figure presents four linked mechanisms. Delegated agency describes how responsibility for task execution is distributed between learner and AI. Epistemic monitoring and verification captures how learners evaluate AI outputs for truthfulness, relevance, adequacy, and fit to task demands. Reconstructive internalization explains how AI-assisted performance is transformed into knowledge and skill that the learner can explain and use independently. Transfer under reduced support indicates that learning is evidenced when the learner can adapt and perform with less reliance on AI. Together these mechanisms distinguish assisted performance from learning and specify when human capability is retained beyond the immediate interaction.
Figure 2: Empirical propositions derived from the mechanisms of Agentivism. This figure maps six empirically testable propositions (P1–P6) onto the four core mechanisms of Agentivism, illustrating how predictions about learning outcomes arise from specific process components.
The four dominant learning theories of behaviorism, cognitivism, constructivism, and connectivism show significant conceptual limitations as generative artificial intelligence (AI) proliferates in educational settings. These frameworks were formulated before the emergence of AI systems capable of generating, synthesizing, and reasoning about knowledge. This article critically examines each learning theory and identifies assumptions challenged by the affordances of generative AI. Drawing on research in distributed cognition, extended mind, human-AI collaboration, AI literacy, cognitive offloading, and metacognition, the article proposes Generativism as a learning theory for the generative AI age. Generativism posits that learning increasingly occurs through the iterative co-construction of knowledge between human learners and AI systems. The proposed framework is organized around four constructs (epistemic partnership, distributed agency, generative literacy, and adaptive metacognition) and four principles specifying the relations among them. The framework offers a foundation for rethinking instructional design, learning, assessment, and expertise development in contexts where generative AI plays an integral role in cognition.
Shan Li, Juan Zheng
Department of Education and Human Services, College of Education, Lehigh University · Department of Community and Global Health, College of Health, Lehigh University
Artificial intelligence (AI) changes social learning when aggregated outputs become training data for future predictions. To study this, we extend the DeGroot model by introducing an AI aggregator that trains on population beliefs and feeds synthesized signals back to agents. We define the learning gap as the deviation of long-run beliefs from the efficient benchmark, allowing us to capture how AI aggregation affects learning. Our main result identifies a threshold in the speed of updating: when the aggregator updates too quickly, there is no positive-measure set of training weights that robustly improves learning across a broad class of environments, whereas such weights exist when updating is sufficiently slow. We then compare global and local architectures. Local aggregators trained on proximate or topic-specific data robustly improve learning in all environments. Consequently, replacing specialized local aggregators with a single global aggregator worsens learning in at least one dimension of the state.
Daron Acemoglu, Tianyi Lin, Asuman Ozdaglar +1
Massachusetts Institute of Technology, NBER, and CEPR · Columbia University · Massachusetts Institute of Technology +1
AI agents pose significant risks as they are granted increasing autonomy. A commonly proposed solution is human oversight and keeping a ''human in the loop'', but this is not a simple solution: Not only do current approaches to AI agent design impede effective human oversight, but the cognitive capacities required for it are also themselves degraded by extended use of AI systems. This position paper argues that current approaches to the development and deployment of AI agent systems do not support effective human oversight -- they contribute to its degradation. To address this, a top priority in the advancement of AI agents should be supporting the situated goals and cognitive requirements of effective human oversight, treating the human needs of overseers at the same level of importance as AI agent capability. To put this idea into practice, we connect work on automation and human-computer interaction to AI agent processes, outlining design-level affordances and organizational protocols that (1) support overseers in exercising critical judgement and (2) counteract the skill atrophy that arises from extended use of automation. We urge developers and deployers to adopt these or similar approaches. Without explicit support for the cognitive demands of effective human-agent interaction, AI agent systems will continue to passively incentivize the degradation of the very human skills they rely on.