Regulation of AI under watch by the EBA

Out-Law News | 21 Jan 2020 | 12:25 pm | 2 min. read

Reforms to financial services regulations could be pursued soon to account for the growing use of artificial intelligence (AI), the European Banking Authority (EBA) has hinted.

The supervisory body said that the development of big data and advanced analytics applications by banks is still "at an early stage" in terms of their sophistication and scope, but it acknowledged the "growing investments and potential opportunities" in this area.

New rules might be needed to account for the "further adoption" of AI and other forms of machine learning (ML) that is anticipated, it said.

"The current trend and pace of change may soon raise the question of the need to develop AI/ML policies or regulatory frameworks for the application of AI/ML in an effort to facilitate its proper development, implementation and adoption within institutions," the EBA said. "The EBA will continue monitoring these developments as part of its mandate on innovation monitoring."

New legislation to support the ethical use of AI is expected to be proposed by the European Commission in the coming weeks, with guidelines in this area already developed for businesses last year. Further reforms in the area of liability for AI have also been mooted. The EBA's comments raise the prospect of regulations being updated in financial services to address firms' use of AI.

The EBA's comments were contained in a new report on big data and advanced analytics in which it said trust in use of those technologies is "essential to allow the utmost benefit from the potential opportunities and at the same time ensure the proper, secure and responsible use of such solutions".

The EBA suggested a new trust framework could be built around eight components – ethics, explainability and interpretability, fairness and avoidance of bias, traceability and auditability, data protection, data quality, security, and consumer protection.

The report, however, noted that there are a number of obstacles to integrating these "elements of trust" facing banks when they are deploying big data and advanced analytics solutions. Two of the challenges it highlighted include those faced by banks in integrating new technology with legacy systems, and with managing increasing volumes of data collected from a variety of sources in an age where there is "increased recognition of citizens’ rights over that data".

The adoption of cloud-based services is helping some banks to overcome the limitations of legacy systems, however, according to the EBA.

"Institutions are increasingly relying on cloud service providers to overcome issues with legacy systems," the EBA said. "Heightened competitive pressure, changing customer behaviour and the speed of new technological releases force institutions to move faster, resulting in an increasing interest in the use of cloud outsourcing solutions in the banking industry. Greater use of [big data and advanced analytics] is perceived to be facilitated by the use of cloud services, which promises high levels of availability, scalability and security."

In its report, the EBA said some of the "elements of trust" it identified could be addressed through a risk-based approach. As an example, it said the level of explainability that banks should have to meet around its use of big data and advanced analytics would depend on "the potential impact on business continuity and/or potential harm to customers" stemming from the use of those technologies.

The EBA highlighted the importance of "human involvement" in the process of decision-making based on advanced analytics-related techniques, and flagged the skills gap that needs to be filled in this area.

"In this new paradigm, solutions are no longer pure IT; new skills in data science are required and a gap has appeared between business and IT experts," the EBA said.

"Staff across an institution may increasingly come to rely on [big data and advanced analytics] applications to support them in their work, suggesting the need for sufficient training in the correct use of such applications to minimise errors and enhance opportunities. Furthermore, all staff and board members could have sufficient understanding of the strengths and limitations of [big data and advanced analytics]-enabled systems," it said.