Recent Advances in Trustworthy Explainable Artificial Intelligence: Status, Challenges, and Perspectives

Abstract

Artificial intelligence (AI) and Machine Learning (ML) have come a long way from the earlier days of conceptual theories, to being an integral part of today's technological society. Rapid growth of AI/ML and their penetration within a plethora of civilian and military applications, while successful, has also opened new challenges and obstacles. With almost no human involvement required for some of the new decision-making AI/ML systems, there is now a pressing need to gain better insights into how these decisions are made. This has given rise to a new field of AI research, Explainable AI (XAI). In this paper, we present a survey of XAI characteristics and properties. We provide an in-depth review of XAI themes, and describe the different methods for designing and developing XAI systems, both during and post model-development. We include a detailed taxonomy of XAI goals, methods, and evaluation, and sketch the major milestones in XAI research. An overview of XAI for security, and cybersecurity of XAI systems, is also provided. Open challenges are delineated, and measures for evaluating XAI system robustness are described.

The Effects of Explanations in Automated Essay Scoring Systems on Student Trust and Motivation

“…For example, to justify outcomes, or to identify what factors play a role, there may be no need to explain the exact inner workings of the system. Moreover, distinctions have been made regarding the method of explanation (Arrieta et al, 2020;Rawal et al, 2022). Explanations can be textual, and visual; can focus on an, example, counterfactual, or simplification (e.g., using a simplified model); can provide feature relevance, knowledge rules, full system descriptions (global explanations), or provide reasoning for a specific decision made by the system (local explanations).…”

The Effects of Explanations in Automated Essay Scoring Systems on Student Trust and Motivation (accepted for publication)

“…For example, to justify outcomes, or to identify what factors play a role in the AI system, there might be no need to explain the exact details of the inner workings of the system. Moreover, distinctions have been made with respect to the method of explanation (Arrieta et al, 2020;Rawal et al, 2021). Explanations can be distinguished in textual and visual explanations, explanations-by-example, counterfactual explanations, explanation by simplification (e.g., using a simplified model), explanation by providing feature relevance, explanation by knowledge rules, full system descriptions (global explanations), or explanations providing reasoning for a specific case or decision being made by the system (local explanations).…”

Artificial Intelligence as a Support Tool to Cybersecurity Activities

“…Elon Musk, Bill Gates, and Stephen Hawking have mentioned that humans need to be alert to the threat of AI. Imagine how dangerous and terrifying it could be if AI-based systems got out of the control of humans [17], [21], [22].…”

“…The main issue is about protecting AI itself against the exploitation of vulnerabilities and other threats caused by targeted attacks on AI [21], [22]. Meanwhile, we should prevent AI from harming humans.…”