Agentivism: New Learning Theory for Human-AI Interaction
A new learning theory called Agentivism has been proposed to address the challenges generative and agentic AI pose to education. The theory argues that when learners delegate cognitive tasks to AI, successful performance no longer guarantees learning. Existing theories like behaviourism, cognitivism, constructivism, and connectivism fail to explain when AI-assisted performance becomes durable human capability. Agentivism defines learning as durable change in the learner, distinct from AI-mediated output. The paper was published on arXiv (ID: 2604.07813) and is relevant to educators, AI developers, and cognitive scientists.
Key facts
- arXiv paper ID: 2604.07813
- Proposes Agentivism as a new learning theory
- Addresses generative and agentic AI in education
- Argues AI-assisted performance may not indicate learning
- Existing theories: behaviourism, cognitivism, constructivism, connectivism
- Agentivism defines learning as durable human capability
- Published on arXiv
- Relevant to AI and education fields
Entities
Institutions
- arXiv