Explainability (XAI)
The effort to make an AI's decisions understandable to humans, so you can see why it produced a given answer.
It matters most in high-stakes uses where the AI said so is not a good enough reason.
Frequently asked questions
What does explainability mean when people talk about AI?
It is the effort to make an AI's decision understandable to a human, so you can see why it produced a given answer rather than just accepting it. In short, it is about turning a black box into something you can question.
Why does explainability matter more in some situations than others?
It matters most in high-stakes uses like lending, hiring, or medical decisions, where the AI said so is not a good enough reason. For low-risk tasks you may not need it, but when the outcome affects someone's life or money, being able to explain the decision is essential.
Can AI always explain how it reached an answer?
Not fully. Modern models are complex, so a clear, complete explanation is often hard to produce, which is the whole reason explainability is an active area of work. Some tools offer partial insight, but a perfect why is not guaranteed.
New to all this? Start with what an AI agent really is, browse the full glossary, or explore the learning hub.
← Back to the glossary