Advancing Trust In AI Algorithms: a State-of-the-Art Examination of Non-Knowledge Aware and Knowledge-Aware Aware Approaches
Ouissem Touameur, Fouzi Harrag · 2023
The use of Artificial Intelligence (AI) systems and large datasets to make decisions on behalf of users is becoming more common in critical areas such as healthcare, agriculture, and law enforcement. However, users are concerned about the trustworthiness of AI algorithms and the sensitive impact of their decisions. To ensure reliable AI systems, several factors must be considered, including explainability, interpretability, transparency, accountability, accuracy, fairness, security, and privacy. This paper presents several state-of-the-art studies aimed at achieving these trust goals. Some papers suggest regulations and laws to control data use and sharing, while others propose methods to identify and mitigate bias. Additionally, the use of Knowledge Graphs (KG) as side information for more accurate and explainable decisions has gained significant interest.