Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Epistemological Fault Lines Between Human and Artificial Intelligence

Walter Quattrociocchi, Valerio Capraro, Matjaž Perc

TL;DR

The paper argues that large language models, despite surface fluency, operate as stochastic pattern-completion systems rather than true epistemic agents. By formalizing text generation as a walk on a high-dimensional linguistic graph and mapping seven stages of human cognition to corresponding AI processes, it uncovers seven epistemic fault lines (grounding, parsing, experience, motivation, causality, metacognition, and value) that sustain a deep divergence between human judgment and AI outputs. It introduces Epistemia as the architectural condition in which linguistic plausibility substitutes for epistemic evaluation, producing the experience of knowing without laborious justification. The authors discuss implications for epistemic evaluation, governance, and literacy, advocating process-aware benchmarks, domain-specific governance, and institutionalized epistemic literacy to preserve accountable judgment in hybrid human–AI systems.

Abstract

Large language models (LLMs) are widely described as artificial intelligence, yet their epistemic profile diverges sharply from human cognition. Here we show that the apparent alignment between human and machine outputs conceals a deeper structural mismatch in how judgments are produced. Tracing the historical shift from symbolic AI and information filtering systems to large-scale generative transformers, we argue that LLMs are not epistemic agents but stochastic pattern-completion systems, formally describable as walks on high-dimensional graphs of linguistic transitions rather than as systems that form beliefs or models of the world. By systematically mapping human and artificial epistemic pipelines, we identify seven epistemic fault lines, divergences in grounding, parsing, experience, motivation, causal reasoning, metacognition, and value. We call the resulting condition Epistemia: a structural situation in which linguistic plausibility substitutes for epistemic evaluation, producing the feeling of knowing without the labor of judgment. We conclude by outlining consequences for evaluation, governance, and epistemic literacy in societies increasingly organized around generative AI.

Epistemological Fault Lines Between Human and Artificial Intelligence

TL;DR

The paper argues that large language models, despite surface fluency, operate as stochastic pattern-completion systems rather than true epistemic agents. By formalizing text generation as a walk on a high-dimensional linguistic graph and mapping seven stages of human cognition to corresponding AI processes, it uncovers seven epistemic fault lines (grounding, parsing, experience, motivation, causality, metacognition, and value) that sustain a deep divergence between human judgment and AI outputs. It introduces Epistemia as the architectural condition in which linguistic plausibility substitutes for epistemic evaluation, producing the experience of knowing without laborious justification. The authors discuss implications for epistemic evaluation, governance, and literacy, advocating process-aware benchmarks, domain-specific governance, and institutionalized epistemic literacy to preserve accountable judgment in hybrid human–AI systems.

Abstract

Large language models (LLMs) are widely described as artificial intelligence, yet their epistemic profile diverges sharply from human cognition. Here we show that the apparent alignment between human and machine outputs conceals a deeper structural mismatch in how judgments are produced. Tracing the historical shift from symbolic AI and information filtering systems to large-scale generative transformers, we argue that LLMs are not epistemic agents but stochastic pattern-completion systems, formally describable as walks on high-dimensional graphs of linguistic transitions rather than as systems that form beliefs or models of the world. By systematically mapping human and artificial epistemic pipelines, we identify seven epistemic fault lines, divergences in grounding, parsing, experience, motivation, causal reasoning, metacognition, and value. We call the resulting condition Epistemia: a structural situation in which linguistic plausibility substitutes for epistemic evaluation, producing the feeling of knowing without the labor of judgment. We conclude by outlining consequences for evaluation, governance, and epistemic literacy in societies increasingly organized around generative AI.
Paper Structure (18 sections, 1 figure, 1 table)

This paper contains 18 sections, 1 figure, 1 table.

Table of Contents

  1. Introduction
  2. Transformers and the Role of Scale
  3. Text Generation as a Walk on a Graph
  4. Human and Artificial Epistemic Pipelines
  5. Sensory and social information vs. Textual input
  6. Perceptual and situational parsing vs. Tokenization and text preprocessing
  7. Memory, intuitions, learned concepts vs. Pattern recognition in embeddings
  8. Emotion, motivation, goals vs. Statistical inference via neural layers
  9. Reasoning and information integration vs. Textual context integration
  10. Metacognitive calibration and error-monitoring vs. Forced confidence and hallucination
  11. Value-sensitive judgment vs. Probabilistic judgment
  12. The Fault Lines
  13. Epistemia
  14. Discussion and Outlook
  15. Epistemic evaluation beyond surface alignment
  16. ...and 3 more sections

Figures (1)

  • Figure 1: The human and LLM epistemic pipelines, each organized into seven corresponding stages.