Explainability in Artificial Intelligence: Techniques, Tools and Applications in Medicine and Bioinformatics

Explainability in artificial intelligence (AI) is essential to increase trust and transparency in AI systems:this study explores not only the techniques and tools developed to make AI models more understandable, but also provides an in-depth analysis of applications in critical fields such as medicine and bioinformatics. Compared to other works on the topic, this study stands out for its focus on the scalability and practical effectiveness of explainability techniques, such as LIME, SHAP, Grad-CAM and DeepLIFT, in specific contexts such as medical diagnostics and genetic analysis. Furthermore, the work discusses the open challenges in the field of XAI, proposing a methodological framework to evaluate and select the most appropriate techniques. The results highlight the need to develop interpretable and scalable models for the ethical and effective use of AI, promoting greater responsibility in the adoption of these technologies in critical fields. In conclusion, this study contributes with a global vision of current progress and future research directions, proposing concrete strategies to improve the practical adoption of XAI.