User Satisfaction and Trust of AI Explanations in Threat Intelligence Analysis: Implications for Acceptance and Usability
Year: 2024
Type: article
Abstract: Threat intelligence analysis is a key component of effective cyber defense, which involves collecting, processing, and analyzing data to identify potential threats and vulnerabilities.Traditional AI based threat intelligence models have play a significant role in combating threats in cyber security domains, however the opaque nature of these "black box" models without explanation on the model decision has created some challenges in responding and managing the increasing complexity of threats in cybersecurity domains.Explainable AI can improve the interpretability and transparency of machine learning models, provides insights into the decisionmaking process, and help users to comprehend the AI decision with the explanation provided leading to better decision making.The paper examines AI explanations and explanation metrics in threat intelligence analysis in cybersecurity operations and how user satisfaction and trust on the explanation output affects acceptance and usability.The paper argues that usefulness, accuracy, clarity, completeness, comprehension, consistency, reliability, timeliness and transparency of the AI explanation can enhance user satisfactions and trust leading to increase acceptance and usability.This study leverages on recent studies on Explainable AI in cybersecurity domain and provides suggestions for enhancing usability and acceptability of AI system based on satisfaction and trust in AI explanations.
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4401814864