Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
Cognitive Redistribution, Not Cognitive Decline: A Critical Review of Studies Linking AI Chatbot Use to Diminished Cognition
0
Zitationen
1
Autoren
2026
Jahr
Abstract
A growing body of research has linked frequent AI chatbot use to reduced neural activation, diminished memory recall, lower critical thinking scores, and increased cognitive offloading. These findings have been widely interpreted—particularly in popular media and anti-AI discourse—as evidence that AI tools are causing cognitive decline. This paper presents a critical review of the key studies driving this narrative, including the MIT Media Lab EEG study (Kosmyna et al., 2025), the Microsoft Research/Carnegie Mellon survey of knowledge workers (Lee et al., 2025), the Gerlich cognitive offloading study (2025), and the Dergaa et al. AICICA theoretical framework (2024). I identify significant methodological limitations across these studies, including small sample sizes, absence of longitudinal data, correlational designs that cannot establish causation, and measurement instruments that conflate reduced effort on offloaded tasks with generalized cognitive impairment. Drawing on Clark and Chalmers’ Extended Mind Thesis (1998), Sparrow et al.’s “Google Effect” research (2011), and 2,400 years of recurring techno-cognitive panic dating to Socrates’ critique of writing in Plato’s Phaedrus, I propose an alternative framework: cognitive redistribution. Rather than diminishing human cognition, AI chatbots appear to be driving a reallocation of cognitive resources—away from tasks that can be effectively offloaded and toward higher-order functions such as evaluation, synthesis, strategic direction, and metacognitive oversight. This pattern is consistent with every prior cognitive technology transition in human history. I argue that what current studies are measuring is the atrophy phase of a cognitive transition, and that interpreting this phase as permanent decline reflects both methodological limitations and a failure to account for the well-documented human capacity for cognitive tool integration.
Ähnliche Arbeiten
Proceedings of the 19th International Joint Conference on Artificial Intelligence
2005 · 5.776 Zit.
An Experiment in Linguistic Synthesis with a Fuzzy Logic Controller
1999 · 5.632 Zit.
An experiment in linguistic synthesis with a fuzzy logic controller
1975 · 5.566 Zit.
A FRAMEWORK FOR REPRESENTING KNOWLEDGE
1988 · 4.548 Zit.
Opinion Paper: “So what if ChatGPT wrote it?” Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy
2023 · 3.372 Zit.