Fig. 3: The AI model life cycle and common biases across each phase.
From: Bias recognition and mitigation strategies in artificial intelligence healthcare applications
This figure maps the stages of the AI model life cycle in healthcare, highlighting the common phase at which biases can be introduced. The AI life cycle is divided into six phases: conception, data collection, pre-processing, in-processing (algorithm development and validation), post-processing (clinical deployment), and post-deployment surveillance. Each phase is prone to specific biases that can affect the fairness, equity, and equality of healthcare delivery.
