Paper Number

ICIS2025-1780

Paper Type

Complete

Abstract

Experts and non-experts increased their usage of Generative Artificial Intelligence (GenAI) and its Large Language Models (LLMs). GenAI is established in companies, organizations, and societies, e.g., supporting grounded decisions and enabling big data analyses. However, misuse risk, inefficient overuse, and discrimination are high. Trustworthy Artificial Intelligence (TAI) requirements and guidelines increase the chances of GenAI and mitigate its challenges. We address chances and challenges through a mixed qualitative and quantitative approach. First, we conducted an integrative literature review of 18 core publications. Second, we interviewed ten experts and analyzed 14 regulatory frameworks. Third, we identify 15 requirements and guidelines and deduce the TrustMap process model. TrustMap comprises four phases for TAI implementation and monitoring, offering a practical, iterative framework that integrates normative principles, regulatory requirements and guidelines, and implementation strategies. Lastly, we summarize our results and findings for theory and practice, discuss our limitations, and deduce a further research agenda.

Comments

12-GenAI

Share

COinS
 
Dec 14th, 12:00 AM

Generative Artificial Intelligence and Large Language Models – A Trustworthiness Process Model

Experts and non-experts increased their usage of Generative Artificial Intelligence (GenAI) and its Large Language Models (LLMs). GenAI is established in companies, organizations, and societies, e.g., supporting grounded decisions and enabling big data analyses. However, misuse risk, inefficient overuse, and discrimination are high. Trustworthy Artificial Intelligence (TAI) requirements and guidelines increase the chances of GenAI and mitigate its challenges. We address chances and challenges through a mixed qualitative and quantitative approach. First, we conducted an integrative literature review of 18 core publications. Second, we interviewed ten experts and analyzed 14 regulatory frameworks. Third, we identify 15 requirements and guidelines and deduce the TrustMap process model. TrustMap comprises four phases for TAI implementation and monitoring, offering a practical, iterative framework that integrates normative principles, regulatory requirements and guidelines, and implementation strategies. Lastly, we summarize our results and findings for theory and practice, discuss our limitations, and deduce a further research agenda.

When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.