UDO MILKAU | Digital Counsellor, European Association of Co-operative Banks (EACB)
The risks associated with the use of artificial intelligence have captured the attention of research, regulation, and industry practitioners in recent years. Given that this is a vast topic in its own right, we are using the experiences of the financial services industry, in specific credit scoring, as a proxy for some of the salient features of AI from a sociotechnical perspective.
Although it shares some of the operational risk challenges associated with other technologies, a model for decision-making reveals how the interfaces with the social context create two new types of risk: naiveté in the use of data for training AI as a statistical classifier and perceptions of the stakeholders regarding its societal implications. While the first can be – and has to be - mitigated by increased literacy within an active internal risk management, the latter requires building trust.