A standalone process for fundamental right impact assessment (hereinafter FRIA) was
also placed among the obligations of the high-risk AI system deployers under Article
27 AI Act. With this enactment, the EU takes one step further to accomplish not only
its international and self-commitments to protect human rights, but also to partly
dissolve the problems of the missing explicit statutory references to ethical principles.
The main objective of this paper is to introduce the FRIA process, place it among
the other risk assessment systems provided by the AI Act. After these general considerations,
the paper aims to compare FRIA with the existing fundamental right impact assessment
methodology in European context, especially with the HUDERIA Methodology.
The last section is dedicated to a specific use case, the automated credit scoring,
where the private financial entities should also conduct a FRIA. The recent EU legislation
on the consumer credit agreements and CJEU judgment on automated decision-making under
GDPR in the Schufa case anticipate that the obligations of providers and deployers
of credit scoring AI systems regarding FRIA will increase in significance.
