L. R

Explainable Artificial Intelligence (XAI): Techniques, Applications, Challenges and Future Directions - A Review

It is concluded that explainability is a necessary, though not sufficient, condition for trustworthy Al, and concrete directions for future research are outlined, including standardised benchmarks, human-centred evaluation, and explainability for large generative models.