EUVIMEDFaigh amach EUVIMED
Open access · CC BY 4.0

LARGE LANGUAGE MODELS AND THEIR CAPACITY TO INDUCE PSYCHOTIC STATES IN THE USER: PROCESS AND ANALYSIS

Anastasiia LAPKO

Epistemological Studies in Philosophy Social and Political Sciences · 2026-05-29 · 9 (1) · 37-44

DOI: 10.15421/342605 ↗

At a glance

Journal article in Epistemological Studies in Philosophy Social and Political Sciences, published online on 2026-05-29. Written by Anastasiia LAPKO. 16 references. The full text is freely available (open access, CC BY 4.0).

AchoimreSonraí leabharliostaCiteRelated studies

Abstract (original language)

This paper explores the epistemological roots of AI-induced psychosis through the lens of Alfred North Whitehead’s process philosophy – the contemporary phenomenon that occurs in the user’s experience during interaction with AI. The purpose of this study is to demonstrate how the theory of symbolic reference and the fallacy of misplaced concreteness provide a robust diagnostic framework for cases where users mistake abstract language models for conscious entities. The author contends that AI-induced psychosis stems from a fundamental categorical mistake: users attribute continuous subjective form and intentionality to discrete tokens, mistaking fluent text generation for genuine sentience. Within the Whiteheadian framework, the user’s proneness to delusions is linked to the fallibility of symbolic reference, specifically when grammatically flawless text on a screen is erroneously endowed with sentience and agency, which Large Language Models are devoid of. The analysis made it possible to identify the following important points: 1. The paper reveals that sycophancy in AI-models reinforces false causal links, transforming symbolic reference into a loop of self-confirming projections. 2. De-abstraction of AI technology through Whitehead’s theories and additional safeguards, proposed by other scholars in the field, will optimistically facilitate “epistemic immunity” development in the user against AI-induced psychosis. 3. By integrating philosophical approaches into digital literacy, a comprehensive model for instrumentalism restoration can be developed based on the distinction between the discrete and the continuous. This paradigm will encourage the user to perceive LLMs as discrete tools rather than continuous “companions”, thereby moderating the risks of AI-induced psychosis and reclaiming the autonomy of human experience in modern digital world.

Cite

APA

LAPKO, A. (2026). LARGE LANGUAGE MODELS AND THEIR CAPACITY TO INDUCE PSYCHOTIC STATES IN THE USER: PROCESS AND ANALYSIS. Epistemological Studies in Philosophy Social and Political Sciences, 9(1), 37-44. https://doi.org/10.15421/342605

Vancouver

LAPKO A. LARGE LANGUAGE MODELS AND THEIR CAPACITY TO INDUCE PSYCHOTIC STATES IN THE USER: PROCESS AND ANALYSIS. Epistemological Studies in Philosophy Social and Political Sciences. 2026;9(1):37-44. doi:10.15421/342605

BibTeX
@article{LAPKO2026,
  author = {LAPKO, Anastasiia},
  title = {LARGE LANGUAGE MODELS AND THEIR CAPACITY TO INDUCE PSYCHOTIC STATES IN THE USER: PROCESS AND ANALYSIS},
  journal = {Epistemological Studies in Philosophy Social and Political Sciences},
  year = {2026},
  volume = {9},
  number = {1},
  pages = {37--44},
  doi = {10.15421/342605}
}
Start for free →