Content area
Full Text
This paper aims to bolster further awareness, and provide additional guidance for banks' corporate governance and national policy-making, to better assess development, deployment and scaling of sophisticated algorithms - Artificial Intelligence (AI). It assesses key considerations and relevant challenges AI brings to the banking industry. It identifies publicly disclosed algorithms and categorizes them in three functional layers - internal, external, and regulatory - applicable to financial services' providers, and compares potential benefits and costs.
It describes various legal tools with differing effects, such as hard regulation, or such as voluntary codes of conduct, which are available to private corporations and national governments alike, and which are aimed at promoting better transparency and legal certainty. The paper also underlies the importance of encouraging innovation while ensuring financial consumers' trust and the adequate protection of their privacy and human rights.
It finally provides a general overview of legal issues which may arise of the use of algorithms by assessing existing investment policies, legislation, and ongoing regulatory developments of various countries such as Canada, US, UK, South Korea and China, while delving into the EU's General Data Protection Regulation's sections dealing specifically with decision-making and profiling applications of algorithms. The paper concludes that although many initiatives have already been undertaken and are ongoing to optimize risk management ofalgorithms, much work still remains, which ought to ultimately peak with a G20 commitment to foundational principles framing the ethical and humane use of AI.
L'auteure invite les banques a accorder plus d'attention a l'intelligence artificielle et propose des guides pour la formulation de regles de gouvernance et de politiques nationales portant sur l'elaboration, le deploiement et la mise a l'échelle d'algorithmes sophistiques. L'auteure souligne les defis poses par l'intelligence artificielle et les principaux facteurs a examiner pour son utilisation dans le secteur bancaire. Elle identifie des algorithmes divulgues publiquement et les classe en trois couches fonctionnelles - interne, externe et reglementaire - applicables aux fournisseurs de services financiers, et compare les benefices escomptes et les couts de chacun.
L'auteure decrit ensuite divers outils, tels l'imposition de reglementations contraignantes ou l'adoption volontaire de codes de conduite, que les entreprises privees et les gouvernements peuvent mettre en place pour promouvoir la transparence et la prévisibilité des solutions juridiques. Elle insiste sur...