AI evidence extraction

Generative AI and the semiosic reconfiguration of knowledge organization – a preliminary exploration

Martin Thellefsen, Bent Sørensen, Amalia Nurma Dewi · Journal of Documentation · 2025

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

9/10

Relevance

1/4

Quality (LMQS)

Evidence

Citations

3.31

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.1108/jd-04-2025-0094

Methodology & findings

Study design

Transdisciplinary semiotic analysis synthesizing theoretical frameworks (Lotman's cultural explosion, Eco's encyclopedic competence, Peircean communication models) to analyze the interpretive complexity introduced by LLMs.

Main result

The paper finds that "the justification for knowledge is shifting from a stable, human-validated 'literary warrant' to a volatile, AI-driven 'algorithmic warrant,'" a change that traditional knowledge organization systems are ill-equipped to handle. The study demonstrates that "LLMs instantiate a dynamic textual environment that challenges conventional KO models built on stability."

Reports effect sizes.

Research paradigm

Interpretivism/Hermeneutics

Author conclusions

"Understanding LLMs as agents of semiosis has implications for the design of future KO systems, suggesting the need for adaptive structures that reflect the fluid, contextualized nature of meaning production in AI-mediated environments." The paper contributes by "integrating Lotman's and Eco's theories with Peircean-informed KO scholarship" to offer "a novel theoretical framework for understanding the epistemological and organizational impact of generative AI technologies."

Open questions raised

The paper identifies the need for methodological development in knowledge organization to accommodate interpretive dynamism and semantic instability introduced by generative AI systems. It emphasizes the requirement for adaptive knowledge organization structures that can handle the fluid, contextualized nature of AI-mediated meaning production.
The authors identify the need for methodological development in knowledge organization to accommodate interpretive dynamism and semantic instability. Future research should develop empirical validation of the theoretical framework and design adaptive KO systems that can handle the fluid, contextualized nature of AI-generated meaning production in AI-mediated environments.
The paper identifies the need for methodological development in knowledge organization to accommodate interpretive dynamism and semantic instability. It opens new avenues for future research in models that can handle the fluid, contextualized nature of meaning production in AI-mediated environments.

Data: not_statedCode: not_statedExtracted from: pdfAgreement 86%

Explore related topics

Epistemic Risk Knowledge Synthesis Conceptual