Abstrakt
One of the most important applications of AI is its use for automated decision-making concerning individuals. This area poses clear risks for fundamental rights as AI decision-making is becoming integrated into everyday life, from assessing loan applications to predicting recidivism, with a potential lack of transparency or biases present in it.
Despite existing legal regulation aimed precisely at this issue - art. 22 GDPR - there is a lack of clear understanding of what constitutes "automated decision-making" in this context. The wording of art. 22 implies a single event to be assessed as either being "solely" based on automated processing or not.
The presentation describes the existing modalities of the use of automated decision-making and the challenge they pose to the binary regulatory view (the art. 22 as well as any future regulatory instrument).