Why Should I Choose You? AutoXAI: A Framework for Selecting and Tuning eXplainable AI Solutions

There are now too many explainability tools to choose from manually. The paper identifies a real problem: as more XAI (explainability) solutions have been developed, organizations struggle to select which one best fits their specific needs without expert guidance.

AutoXAI automates the selection process by matching explainability tools to your actual constraints. Rather than requiring technical expertise to evaluate dozens of options, the framework considers your specific context—including your dataset type, the ML model being used, and your particular explainability needs—to recommend the most appropriate tool and its settings.

Different XAI solutions work better for different situations. The paper demonstrates through two use cases that there is no one-size-fits-all explainability solution; the best choice depends on factors like your dataset characteristics and what you're trying to explain, making automated recommendation valuable for responsible AI practitioners.

This framework borrows proven techniques from other AI fields. AutoXAI adapts methods from recommender systems and automated machine learning (AutoML), applying established optimization strategies to solve the XAI selection problem rather than starting from scratch.

Selecting the right explainability tool matters for responsible AI deployment. By making it easier for organizations to find well-matched XAI solutions with optimized settings, this framework helps ensure that AI systems can be properly explained to stakeholders—a key requirement for trustworthy and accountable AI systems.