Neural Interaction Between Learners and Artificial Intelligence Through the Lens of Fallacies

The rapid expansion of large language models and generative artificial intelligence systems in educational ecosystems has moved learner– machine interaction beyond a merely instrumental interface toward cognitive entanglement and neural stimulation. Adopting a phenomenological and critical approach, this qualitative study explores learners’ neural–cognitive engagement with fabricated content arising from “AI hallucination” and its implications for the reproduction of epistemological fallacies. The research involved in-depth semi-structured interviews with higher education learners and participant observation of AI-assisted problem-solving processes. The resulting data were interpreted through inductive thematic analysis. The findings indicate that, when confronted with algorithmic error, learners’ cognitive processing develops an “artificial confidence bias” under the influence of AI’s coherent linguistic structure and persuasive tone. By providing rapid access to seemingly complete answers, this process satisfies the neural dopaminergic reward circuit and consequently suspends cognitive inhibition and critical reflection. The extracted themes revealed that learners’ neural interaction with generative AI hallucinations, through six fallacies— formal eloquence, artificial authority, immediate relief, part-to-whole generalization, anthropomorphic analogy, and confirmatory conformity— transforms human agency into cognitive passivity and erodes critical judgment through artificial confidence bias. This epistemic erosion reduces truth to statistical rhetoric and can be mitigated and remedied only through a transition from reward-driven consumption to reflective pedagogy and relational ethics.