The development and use of new technologies, based on algorithms and artificial intelligence, have opened new frontiers about scoring systems and, in general, creditworthiness assessment. In Italy, the traditional credit scoring systems of the banks and other financial institutions are based on linear regression models and not yet on artificial intelligence and machine learning applications to identify complex relationships between data. The proposal of a European Regulation laying down harmonised rules on artificial intelligence (Artificial Intelligence Act) provides that AI credit scoring systems should be classified as high-risk AI systems: they may lead to discrimination of persons or groups based, for instance, on racial or ethnic origins or create new forms of discriminatory impacts. The essay discusses the regulatory issues to address so that credit scoring systems are not discriminatory but, at the same time, do not compromise compliance with the rules of sound and prudent management of financial institutions.
Credit scoring 5.0, tra Artificial Intelligence Act e Testo Unico Bancario / G.L. Greco. - In: RIVISTA TRIMESTRALE DI DIRITTO DELL’ECONOMIA. - ISSN 2036-4873. - 2021:3 suppl.(2021), pp. 74-100.
Credit scoring 5.0, tra Artificial Intelligence Act e Testo Unico Bancario
G.L. Greco
2021
Abstract
The development and use of new technologies, based on algorithms and artificial intelligence, have opened new frontiers about scoring systems and, in general, creditworthiness assessment. In Italy, the traditional credit scoring systems of the banks and other financial institutions are based on linear regression models and not yet on artificial intelligence and machine learning applications to identify complex relationships between data. The proposal of a European Regulation laying down harmonised rules on artificial intelligence (Artificial Intelligence Act) provides that AI credit scoring systems should be classified as high-risk AI systems: they may lead to discrimination of persons or groups based, for instance, on racial or ethnic origins or create new forms of discriminatory impacts. The essay discusses the regulatory issues to address so that credit scoring systems are not discriminatory but, at the same time, do not compromise compliance with the rules of sound and prudent management of financial institutions.File | Dimensione | Formato | |
---|---|---|---|
2021 Greco credit scoring 5.0 artificial intelligence act e tub.pdf
accesso aperto
Tipologia:
Publisher's version/PDF
Dimensione
889.8 kB
Formato
Adobe PDF
|
889.8 kB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.