Trade-offs in Financial AI: Explainability in a Trilemma with Accuracy and Compliance

Accuracy and compliance are non-negotiable prerequisites, not trade-offs: The study found that finance professionals view accuracy and regulatory compliance as "hygiene factors" that must be met first—without them, an AI system is considered a liability regardless of how explainable it is, fundamentally challenging the common assumption that explainability requires sacrificing accuracy.

Explainability is a "gateway to adoption" rather than a primary goal: Rather than being pursued for its own sake, explainability functions as a trust mechanism that determines whether AI tools are actually used and defended in practice, suggesting that transparency matters most when it enables human confidence and organizational buy-in.

Financial AI operates within a complex system of constraints, not simple trade-offs: The research reveals that financial institutions must simultaneously navigate regulatory requirements, budget limitations, and speed demands alongside accuracy and explainability—creating a "trilemma" of competing priorities that cannot be reduced to simple binary choices between accuracy and transparency.

Operational factors (cost and speed) act as secondary gatekeepers: Even when accuracy, compliance, and explainability are achieved, practical feasibility constraints around implementation costs and processing speed determine whether an AI system can actually be deployed, indicating that responsible AI requires attention to real-world operational realities.

One-size-fits-all XAI solutions are insufficient for finance: The study's inclusion of perspectives from C-suite executives, developers, and regulators across multiple regions suggests that different stakeholders have distinct priorities, implying that effective responsible AI implementation requires tailored approaches rather than universal explainability standards.