Algorithmic Impact Assessment

A systematic process to evaluate the potential social, ethical, economic and legal impacts of deploying an algorithmic system before and during its use.

Before launching an AI system, you study what could go wrong and who might be affected. Similar to an environmental impact study before building a factory, but for technology.

Impact assessments are a core component of AI governance. Many regulatory frameworks either require or strongly recommend them before deploying AI systems. Establishing a consistent impact assessment process across your organisation is a governance priority.