AI ethics in the context of financial services
The Alan Turing Institute and the Financial Conduct Authority are collaborating on a year-long project around the use of artificial intelligence in the financial services sector. Here, Dr Florian Ostmann, Policy Fellow at The Alan Turing Institute, comments on the motivation for the project and the Turing Institute’s work on AI governance more generally.
Advances in data science and artificial intelligence are set to transform business processes throughout the economy. The rapidly growing range of potential applications of AI in business holds the promise of wide-ranging benefits to consumers, companies, and society at large.
Yet, there is also the possibility of significant negative consequences, with some applications giving rise to important ethical questions. Considering this combination of potential benefits and risks, responsible conduct and ethical due diligence is fundamental to the sustainable adoption of AI technologies.
When adopting an industry-specific lens on AI governance, financial services represents an important area of consideration. Applications in this sector include use cases with high stakes, both in terms of benefits and risks. In banking, the use of advanced analytics and novel algorithmic tools has the potential to make loans and other services of existential significance more accessible and affordable, for example. In the absence of effective safeguards, however, reliance on algorithmic tools could also have negative implications such as the effective exclusion of certain customer segments from the market or the amplification of existing biases or patterns of discrimination.
Pursuing research on AI governance that evolves in lockstep with technological advances is a cornerstone of The Alan Turing Institute’s work. In this context, AI in financial services constitutes an important focal point.
Earlier this year, we convened a two-day conference on AI ethics in the financial sector that brought together financial service professionals, regulators, consumer advocates and a diverse range of academic researchers. At the same time, we are partnering with the Financial Conduct Authority on a year-long project, examining current and anticipated uses of AI in financial services, analysing the ethical and regulatory questions that arise in this area and exploring potential strategies for addressing them.
Culminating in a report to be published in 2020, this joint project places a special focus on transparency and explainability. While not exhaustive of the diverse range of ethical questions that arise in the context of AI, these considerations play a key role, both as a matter of general trustworthiness and as a pre-condition for addressing other substantive concerns. In examining them, the project aligns with several other ongoing projects at the Institute, such as our collaboration with the Information Commissioner’s Office on developing a framework for explaining algorithmic decision-making.
The Alan Turing Institute is the national institute for data science and artificial intelligence, with headquarters at the British Library.