Explainable AI for Non-Expert End Users: A Literature Review

Resumo

The exponential growth of Artificial Intelligence (AI) systems in everyday applications such as healthcare, finance, education, and public services has increased the need for transparency and user trust. Explainable Artificial Intelligence (XAI) aims to make AI decisions understandable to humans; however, much of the existing research focuses on technical interpretability methods rather than the needs of non-expert end users. The paper presents a systematic literature review of user-centred XAI approaches designed for individuals without specialised AI knowledge. The review examines the recent peer-reviewed studies to recognize frequent explanation techniques, interface designs, evaluation methodologies, and application domains that prioritizes human comprehension. Moreover, the study categorises explanation formats such as visual, textual, rule-based, and interactive explanations and inspects their effectiveness in enhancing trust, usability, and decision confidence among lay users. Key challenges are highlighted, which include cognitive overload, lack of standardized evaluation metrics, accessibility barriers, and domain-specific biases. The study also displays a research gap in personalization, cross-cultural usability, and long-term user engagement. In future research goals are proposed, highlighting adaptive explanation systems, multimodal interaction, and inclusive design principles. The study aims to provide researchers and practitioners with a synthesized understanding of the human-centred XAI and to support the development of transparent AI systems that are accessible and meaningful to the general public.