TY  - JOUR
AU  - Lamadé, Annegret
AU  - Beekmann, Dustin
AU  - Eickhoff, Simon
AU  - Grefkes, Christian
AU  - Tscherpel, Caroline
AU  - Meyding-Lamadé, Uta
AU  - Bassa, Burc
TI  - Qualitätsindikatoren künstliche Intelligenz - Quality indicators artificial intelligence
JO  - Der Nervenarzt
VL  - 95
SN  - 0028-2804
CY  - Heidelberg
PB  - Springer
M1  - FZJ-2024-01161
SP  - 242-246
PY  - 2024
N1  - Kein Post-print verfügbar
AB  - The ability of some artificial intelligence (AI) systems to autonomously evolve and the sometimes very limited possibilities to comprehend their decision-making processes present new challenges to our legal system. At a European level this has led to reform efforts, of which the proposal for a European AI regulation promises to close regulatory gaps in existing product safety law through cross-sectoral AI-specific safety requirements. A prerequisite, however, would be that the EU legislator does not only avoid duplications and contradictions with existing safety requirements but also refrains from imposing exaggerated and unattainable demands. If this were to be taken into consideration, the new safety requirements could also be used to specify the undefined standard of care in liability law. Nevertheless, challenges in the context of provability continue to remain unresolved, posing a risk of rendering the legal protection efforts of the aggrieved party ineffective. It remains to be seen whether the EU legislator will address this need for reform with the recently proposed reform of product liability law by the Commission.
LB  - PUB:(DE-HGF)16
C6  - 38085285
UR  - <Go to ISI:>//WOS:001127314200003
DO  - DOI:10.1007/s00115-023-01573-6
UR  - https://juser.fz-juelich.de/record/1022024
ER  -