Explainability means people can understand why an AI reached a result.
Example
A rejected loan application names the main reasons, such as income too low relative to the instalment.
For pros
Methods range from simple, explainable models to after-the-fact explanations (e.g. feature importance). For large language models explanations are difficult; reasons the model writes itself need not be accurate. The EU AI Act gives affected people a right to an explanation for certain high-risk decisions (Art. 86).
