Excessive transparency with regard to AI has downside, Singaporean regulator warns
MAS says excessive transparency could create confusion or unintended opportunities for individuals to exploit or manipulate AIDA models.
In line with plans unveiled earlier this year, the Monetary Authority of Singapore (MAS) has earlier today outlined the Principles for the use of artificial intelligence and data analytics (AIDA) in decision-making in the provision of financial products and services.
Among these key principles is that of transparency, with the regulator noting that a balance should be struck in this respect.
Whereas increased transparency in AIDA firms’ use of AIDA could improve public understanding and confidence in AIDA, excessive transparency could create confusion or unintended opportunities for individuals to exploit or manipulate AIDA models, MAS says. It is important to balance these considerations in determining the appropriate level of transparency in the use of AIDA. In determining levels of transparency, the materiality of the decision is also relevant.
Specifically, data subjects may seek clear explanations about data used for decision-making, how data affects the decision and the consequences on data subjects. Clear explanations, however, do not necessitate exposure of intellectual property or publishing of proprietary source codes. Clear explanations could instead focus on facilitating the data subject’s understanding of the firms’ use of AIDA.
For instance, let’s say a firm uses AIDA models for fraud detection and to identify possible “red flags”. Given materiality considerations and concerns of model manipulation or exploitation, this firm may decide not to share about the AIDA model used, or provide explanations relating to this area. On the other hand, if the firm deploys an AI phone operator to handle customer calls and engage directly with customers, it should opt to begin the conversation by informing the data subject that they are interacting with an AI phone operator.
The rest of the principles focus on Fairness, Accuracy, and Accountability. For instance, firms have to make sure that individuals are not systematically disadvantaged through AIDA-driven decisions unless these decisions can be justified.
This set of Principles is not intended to be prescriptive. Through the process of ongoing industry engagement, there is scope for MAS to identify areas where more specific or technical guidance would benefit the industry, with this set of Principles serving as a foundational framework for consideration and discussion.