FinTrust: Trust Engineering for the Financial Industry

The FinTech industry is one of the major growth industries in the United Kingdom. These companies create new, cheaper and faster services, utilizing the latest technologies such as cloud, mobile and blockchain. To succeed they need to gain the trust of customers in a period that society's trust in the financial industry is still impacted by the mortgage crisis almost a decade ago. They need to gain this trust while technologies are changing rapidly and data breaches are continuously in the news. FinTrust will research the issue of trust in FinTech, identifying the generic research challenges and establishing fundamental research results. A particular focus will be on increased automation through the use of machine learning algorithms, which may have implications that affect consumer trust in the new services.

For the first time, research results as derived in FinTrust will be made available as a Trust Engineering Tool Kit, allowing service provider, regulators and consumer organizations to assess levels of trust, quantify the extent to which trust matters to consumers and trust attitude differentiates consumers, assess bias in automated advice, establish whether algorithms may lead to financial exclusion or financial distress and protect services from erroneous algorithmic results. In our work we balance the commercial interest with society's interests, addressing issues such as financial exclusion through algorithmic bias.

FinTrust offers a focused pathway to academic, societal and economic impact. Starting from the co-creation effort with Atom Bank that resulted in the proposal, we grow the Stakeholder Panel to include customer organizations, regulators and other stakeholders. We then establish a joint-up academic community in a FinTech Network, bringing together ICT, EPSRC and other research and intertwine this with existing research networks in finance and business research. We will make the software tools available freely and to a wide audience, training the research staff in interacting and public engagement to optimize the impact from our research.

The research in FinTrust aims to derive software and management tools for trust engineering in the financial industry, especially focusing on automation through machine learning algorithms . FinTrust delivers a tool kit for assessing human trust, assessing bias of algorithms, and offers approaches to control how socio-technical systems interface with algorithms. Our tools are based on FinTrust's generic research insights and can therefore find application across the industry (even beyond the FinTech industry, although initially this is not the focus of the project and its pathways to impact). The tools will be open source, and downloadable from the web and the data we use will confirm to the policy framework on research data of UK Research Councils. That is, we make all data openly available as quickly and with as few restrictions as possible, realising the commercial constraints that may be posed by our partner Atom Bank.

When co-creating the FinTrust proposal we worked closely with Atom Bank, one of the most prominent FinTech companies in the UK. In the six months of FinTrust we will establish a Stakeholder Panel to broaden the base of interaction, including FinTech companies, regulatory bodies (Financial Conduct Authority) and consumer protection organisations. The stakeholders will collaborate through further co-design of the research approaches, as well as in co-research through participation in the research in human factors and in testing our tools. In this manner, FinTrust will optimize the suitability of its tool for industrial take-up and for functionality that protect the consumer while fulfilling requirements of the regulator.

A specific area of interest in our work is the issue of financial exclusion and financial distress. These have been highlighted as important societal challenges by the House of Lords ('Tackling Financial Exclusion: A country that works for everyone? Report of Session 2016-17', House of Lords Select Committee on Financial Exclusion, 2017) and the Financial Conduct Authority ('Predicting financial distress by predicting unaffordable consumer credit agreements: An applied framework', Financial Conduct Authority Occasional Paper 28, 2017). By interfacing with FCO and consumer organisations we can optimize the tools we develop in terms of avoiding bad loans and, at the same time, financial exclusion.

FinTrust will deliver benefits in terms of raising public awareness over the operations and implications of banking services, including fully automated 'robo-advice'. It will also reach out and discuss with consumers, politicians and other about the implications of automated advice for financial exclusion, distress and consumer trust.

The project will pay particularly close attention to career development if its research associates, through public engagement training, training sessions with our partner Atom Bank, and through involving them intensively with the various network activities FinTrust leads and participates in. We will develop tutorials and presentations for academic and non-academic conferences and meetings, reserving budget for professional developers of presentations materials, flyers and web site.