Money2020: financial services must 'own' ethical use of AI

Out-Law Analysis | 04 Jun 2019 | 9:26 am | 2 min. read

There is growing recognition of the importance of ethical use of artificial intelligence (AI), but there remain challenges for financial services institutions over how to operationalise guidelines that have been developed so far and which function of the business should 'own' such a project.

AI was the hot topic of discussion on the first day of the Money20/20 Europe 2019 conference in Amsterdam. The conference is Europe's largest fintech event and is meant to "enable payments and financial services innovation for connected commerce at the intersection of mobile, retail, marketing services, data and technology". Pinsent Masons is a sponsor of the event.

AI is a badge applied to a wide range of capabilities; process automation and logic-based processing most commonly, but machine learning is big growth area now we have easy access to computing power. AI, in this sense, is still rare, but in the future we might see development beyond today's chatbot and fraud prevention solutions to AI that takes care of the entire customer journey, from background process tasks to pricing and complex interaction.

Whatever AI is said to mean currently, research carried out by Gartner and shared at the conference has confirmed its move to becoming 'business as usual' technology.

According to Gartner, AI will be in 37% of organisations before the end of this year. Use of AI in financial services is likely to be greater than this given how well disposed the sector is to such technology.

...whilst 50% of chief executives see AI as the most important technology for their organisation, that view is shared by just 4% of chief administrative officers

Yet the research confirmed that the importance of AI to financial services companies is rated differently within the boardroom: whilst 50% of chief executives see AI as the most important technology for their organisation, that view is shared by just 4% of chief administrative officers.

The statistics show that more work is needed to raise awareness of how AI can improve the way financial institutions perform.

A central barrier to use of AI in the sector is safe adoption that avoids ethical or discriminatory issues - this is particularly important as businesses look to scale with a platform mentality so that what might be a small glitch early on compounds and causes increasing risk if not identified and checked.

Leadership from global standards bodies is needed to help take forward work already done in this area.

Earlier this year, a high-level expert group on artificial intelligence (AI HLEG), set up by the European Commission, published guidelines for trustworthy AI, while similar principles were also adopted by the OECD's 36 member countries, along with Argentina, Brazil, Colombia, Costa Rica, Peru and Romania last month. Ethical issues are also being explored by the UK's newly created Centre for Data Ethics and Innovation and Office for AI.

The guidelines, principles and recommendations developed to-date are necessary, as without confidence in the ethical and trustworthy use of data and AI, the economic and societal benefits promised from utilising the technology will not be achieved and innovation will suffer. However, they only go so far.

Organisations exploring the use of AI in financial services need the help of global standards bodies, such as IOSCO and the Financial Stability Board, to shape more detailed requirements on the use of AI that address the issue of ethics, as well as legal and regulatory compliance.

As financial services institutions look to develop AI-driven products and services, however, there is a further need to ensure that there is appropriate ownership internally of any project to operationalise ethical guidelines.

Some institutions may look to the legal function to take this responsibility on, but this may not be a wholly natural fit. In a post-financial crisis age, where institutions have more distinct business and control functions, it is vital that there is ownership of ethical use of AI at all stages of design, development, implementation and review of such technology.

As Pinsent Masons has found, in research carried out in partnership with Innovate Finance, most UK customers are open to the use of AI in financial services. Getting ethics right is an important step towards winning the trust of naysayers and delivering the full benefits the technology offers for both business and consumers.

Angus McFadyen and Luke Scanlon are experts in financial services and technology law at Pinsent Masons. Pinsent Masons is a sponsor of the Money20/20 Europe 2019 conference in Amsterdam, Europe's largest fintech event.