Mapping Trustworthiness in Large Language Models: A Bibliometric Analysis Bridging Theory to Practice
José Siqueira de Cerqueira, Kai-Kristian Kemell, Rebekah Rousi, Nannan Xi, Juho Hamari, Pekka Abrahamsson
TL;DR
This paper addresses the lack of consensus on defining and operationalizing trustworthiness in large language models (LLMs) and aims to bridge theory and practice through a two-stage analysis. It conducts a bibliometric mapping of 2,006 Web of Science publications from 2019–2025 using Bibliometrix, complemented by a manual review of 68 papers that explicitly discuss trust, to extract definitions and practical techniques across the LLM lifecycle. The study finds 18 distinct definitions of trust/trustworthiness and 20 practical strategies, with a heavy emphasis on post-training, developer-driven approaches such as supervised fine-tuning and retrieval-augmented generation (RAG); it also highlights fragmentation and risks of ethics washing, urging standardized frameworks and stronger regulation. Overall, the work provides a foundation for researchers, developers, and policymakers to move toward unified terminology, governance, and actionable guidance for trustworthy deployment of genAI systems.
Abstract
The rapid proliferation of Large Language Models (LLMs) has raised significant trustworthiness and ethical concerns. Despite the widespread adoption of LLMs across domains, there is still no clear consensus on how to define and operationalise trustworthiness. This study aims to bridge the gap between theoretical discussion and practical implementation by analysing research trends, definitions of trustworthiness, and practical techniques. We conducted a bibliometric mapping analysis of 2,006 publications from Web of Science (2019-2025) using the Bibliometrix, and manually reviewed 68 papers. We found a shift from traditional AI ethics discussion to LLM trustworthiness frameworks. We identified 18 different definitions of trust/trustworthiness, with transparency, explainability and reliability emerging as the most common dimensions. We identified 20 strategies to enhance LLM trustworthiness, with fine-tuning and retrieval-augmented generation (RAG) being the most prominent. Most of the strategies are developer-driven and applied during the post-training phase. Several authors propose fragmented terminologies rather than unified frameworks, leading to the risks of "ethics washing," where ethical discourse is adopted without a genuine regulatory commitment. Our findings highlight: persistent gaps between theoretical taxonomies and practical implementation, the crucial role of the developer in operationalising trust, and call for standardised frameworks and stronger regulatory measures to enable trustworthy and ethical deployment of LLMs.