Where appropriate, users, impacted third parties, and actors in the AI life cycle should have the ability to contest an AI decision or outcome that is harmful or creates adverse effects. It should be ensured that impacted parties have the necessary information to contest outcomes with AI deployers and AI developers using these routes, such as information indicating that a decision was informed by an AI System.
The importance of transparency and explainability in enabling contestability should be communicated effectively. Methods should be developed to enable users to understand and utilise available routes to contest AI-informed decisions.
|