Mario López Ayala, PhD
The next digital divide is epistemic.
Artificial intelligence is no longer merely helping people find information. It increasingly influences which questions they ask and which explanations appear credible. The central issue is no longer only what AI knows, but how its presence changes what humans accept as knowledge.
People rarely surrender judgment through a single conscious decision. Delegation develops through convenience: an algorithm responds faster and reduces the effort needed to compare alternatives. With repetition, assistance can acquire the appearance of authority, although fluency is not evidence and confidence is not truth.
Human responses to AI are neither uniformly trusting nor resistant; they vary with the task, context, perceived accuracy and authority attributed to the system [1]. Experiments indicate that repeated interaction with biased AI can alter human judgments, creating feedback loops through which errors are internalized and amplified [2]. The danger is not simply that an algorithm may be wrong, but that its error may enter human reasoning.
Human–AI collaboration is not automatically superior. A systematic review and meta-analysis found that combined systems improved human performance but did not consistently outperform the better of the human or AI working independently [3]. The decisive question is not which side is universally better, but when their interaction strengthens—or weakens—judgment.
The consequences extend beyond individual users. In education, students may replace exploration with immediate synthesis; in journalism, automated systems can accelerate verification while producing plausible falsehoods at scale. UNESCO therefore emphasizes human agency, critical thinking and responsible AI competencies in education and research [4].
Digital literacy alone is no longer enough. Operating a platform does not prepare someone to question why an answer appeared, what evidence supports it or whose assumptions shaped the system. Societies increasingly need epistemic literacy: the capacity to evaluate how knowledge is produced, filtered, legitimized and contested in human–machine environments.
This premise guides a research program I define as Human–AI Epistemic Behavior. My proposed multidimensional framework examines how trust, worldview, digital literacy, perceived authority, institutional context and algorithmic design interact when people form beliefs or make decisions with AI. Consistent with sociotechnical approaches emphasizing trustworthiness, context and human responsibility [5], it treats judgment as dynamic and configurational—effects emerge from combinations of factors rather than a single cause. It also proposes that the same AI response may be accepted, questioned or rejected differently across cultures because trust is filtered through experience, institutional legitimacy and worldview.
The framework remains a theory-building proposal requiring empirical testing across cultures, institutions, age groups and forms of human–AI interaction. Yet the next digital divide may separate those who use AI while preserving their capacity to question from those who outsource that capacity without noticing. Artificial intelligence may help humanity know more than ever—but the future of knowledge will depend on whether humans remain responsible for deciding what deserves to be believed.
Selected References
[1] Jussupow, E., Benbasat, I., & Heinzl, A. (2024). An integrative perspective on algorithm aversion and appreciation in decision-making. MIS Quarterly, 48(4), 1575–1590. DOI: 10.25300/MISQ/2024/18512.
[2] Glickman, M., & Sharot, T. (2025). How human–AI feedback loops alter human perceptual, emotional and social judgements. Nature Human Behaviour, 9, 345–359. DOI: 10.1038/s41562-024-02077-2.
[3] Vaccaro, M., Almaatouq, A., & Malone, T. (2024). When combinations of humans and AI are useful: A systematic review and meta-analysis. Nature Human Behaviour, 8, 2293–2303. DOI: 10.1038/s41562-024-02024-1.
[4] Miao, F., & Holmes, W. (2023). Guidance for generative AI in education and research. UNESCO.
[5] National Institute of Standards and Technology. (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0). NIST AI 100-1.
This article draws on the author’s ongoing research into Human–AI Epistemic Behavior and the emerging ways artificial intelligence influences human judgment, trust and knowledge formation.