A fundamental rights impact assessment is an analysis that examines how the deployment of an AI system may affect the rights of the people whom its decisions concern. It differs from an ordinary risk assessment in its perspective: it does not evaluate risk to the company, but to the job applicant, client, or citizen. It typically describes the purpose and context of use, the categories of affected persons, which decisions the system influences, what risk of discrimination or incorrect evaluation exists, how that risk is mitigated, and how a person can challenge the outcome. It is related to a data protection impact assessment but does not replace it – one addresses privacy, the other broader rights. It is carried out before deployment and updated on any substantial change. Discuss the specific scope of the obligation for your company with a lawyer.
See also: High-risk AI system, Human oversight of AI, AI risk classification.
