In the context of AI, transparency refers to the openness and accessibility of the "inner workings" of an AI system, requiring easy-to-understand explanations of algorithmic models, the data driving them, and the rationale for their use. It also encompasses transparency in the use of AI systems by public or private actors, especially toward individuals likely to be impacted by these systems. Transparency aims to help people develop a basic understanding of AI systems, including the reasons for their development, training, deployment, and impact.

The principle of transparency is closely linked to the concept of explainability in AI ethics, which calls for providing information that a non-technical person can understand, particularly regarding the processes driving AI outcomes.

As principles of responsible AI governance, transparency and explainability require that information about processes, practices, and decisions be open, accessible, and easily understood by a broad range of stakeholders, including government entities, private sector companies, civil society organizations, and academic institutions.