In the second article of his explainable AI series, Paul Larsen looks at what today’s XAI tools really deliver for different stakeholders—from users to regulators—and where they still fall short for trust, liability and high-risk decisions.