Year: 2025
Type: conference-paper
Abstract: Informācija (mākslīgais intelekts) pret Ļaunumu. Ukrainas kara kontekstsInformation (artificial intelligence) Against Evil. The Ukrainian War ContextThis work is a seminar presentation discussing the role of information and artificial intelligence (AI) in the context of the Ukrainian war. The presentation makes the following arguments:AI can be a powerful tool for defending truth and exposing disinformation. Complex information can help individuals resist propaganda and evil. There is an information war happening alongside the physical war in Ukraine. The author proposes using books as "cognitive detectors of moral alignment" and suggests that after the war, Russian soldiers should be forced to study these books.The document says that there are many conversations with LLMs about the war in Ukraine, and these conversations can be used to examine the different levels of information complexity. The levels of complexity can be used to determine if LLMs can be used against evil, since evil usually lacks logical backgrounds.The document mentions that LLM responses are used in the development of an AI-assisted semantic evaluation framework. In this framework: Each book in the set becomes a test prompt for the LLM. The LLM is asked a series of structured questions related to each book. Humans are then asked to paraphrase or interpret these questions and the LLM's answers. The results are evaluated for depth, accuracy, and alignment with truth and moral reasoning. So, the LLM's responses are used as a basis for further evaluation of both the LLM's understanding and how humans interpret its responses.
Author: Zeps, Dainis
Institution: University of Latvia
Language: English
Subfield: Sociology and Political Science
Field: Social Sciences
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W6911865464
FWCI:
Cites:
Cited by:
Related to: