AI Impact Assessments (AIIA)

AI Impact Assessments

An impact assessment ensures an organization has sufficiently considered a system’s relative benefits, risks, and costs before implementation. An AI Impact Assessment (AIIA) is utilized to facilitate discussions and documentation on AI systems’ use cases. It explores obstacles in the data, the system, and the algorithms, taking account of applicable rules and regulations. An AIIA serves as an instrument for dialogue and for recording thought processes, enhancing accountability, quality, and reproducibility of an AI system.

The commissioning client bears primary responsibility for implementing or commissioning the AIIA. An AIIA must be drawn up for each use case of the AI system. However, each AIIA is expressly completed in a proportionate manner, befitting the impact and the risk profile of the application. An organization might use an impact assessment to understand how an AI system will impact users and society more broadly at every stage: data gathering, development, deployment, and continuous monitoring in society. These assessments can serve as critical guardrails for public and private organizations, with the investment required proportionate to risk. For instance, AI systems that offer movie recommendations would receive less scrutiny than those that make hiring or loan decisions. An impact assessment promotes accountability by requiring an organization to document its decision-making process.