INFORMATIONAL SUSTAINABILITY AND GENERATIVE ARTIFICIAL INTELLIGENCE: EPISTEMIC LIMITS OF KNOWLEDGE AUTOMATION
LIMITES EPISTÊMICOS
Keywords:
Artificial Intelligence, Knowledge Management, Informational SustainabilityAbstract
The present article critically analyzes the concept of informational sustainability in light of the growing use of artificial-intelligence–based language models for knowledge representation. It argues that, because these systems operate by probabilistic inference, they can strain classical principles of Knowledge Organization, such as semantic consistency, terminological control, and classificatory transparency. Based on a theoretical-conceptual and bibliographic approach, it offers a theoretical essay and discusses how algorithmic opacity can undermine the reliability, justifiability, and interpretive preservation of produced knowledge. It is argued that informational sustainability is not limited to data preservation but involves maintaining coherent interpretive structures over time. The article concludes that incorporating AI requires revising epistemological criteria in Knowledge Management through human curation, otherwise there is a risk of weakening informational mediation and amplifying cognitive and cultural biases