A few months ago, one of the most established banking businesses n the world – Lloyds Bank, confirmed it was joining the likes of Google and Uber by delivering a new role – Group Head of Data Ethics. The new position is mainly responsible for the moral and legal obligations for collecting, measuring and protecting significant volumes of data available at Lloyds Bank.
The job posting focused on promoting, embedding and commercialising data and analytics and the culture within the business to drive a data-enabled organisation. If data is the focus to drive business insights, AI is the accelerator that will power this process forward, generating new solutions quicker than previously possible. Conversely, using AI automation on scale without a data ethics representative in a highly regulated industry like finance is similar to beginning a journey without having a map, lacking direction or the ability to shift their course.
The pressure of audits, compliance checks, detailed processes and meeting standards in the financial sector all influence overall success, not just in terms of avoiding regulatory fines but reducing any possible legal implications due to data errors. The challenge comes when the rate of data creation exceeds the ability to process it efficiently.
To appreciate why data ethics is so important, we need to understand the role data has in the finance sector. Combining the complexity and volume of data created today, this surge of information rapidly outpaced the original systems and processes needed to analyse data efficiently. Traditionally, compliance and regulatory processes were managed manually via static spreadsheets. Today, these processes are being streamlined through advanced analytics, making the technologies more accessible and approachable. Analytical automation is a solution to refine large volumes of data effectively and convert it into clear, actionable insights.
Two main tools used within the finance industry are business process automation and automated insight generation from data. Process automation is the ability to take a selected activity, such as extracting data from its source and combining it with data from other areas, and automatically delivering reports. This process generates several benefits for compliance and regulatory requirements and significantly reduces report generation with nearly no errors. These results happen through repeatable, transparent and verified processes completed the same way every time.
The second tool is applying AI for decision-making. AI can determine patterns within data sets to identify fraud or money laundering of specific factors, impacting the ability of an applicant to repay their mortgage. The speed and accuracy benefits that come with automation mean this technology is quickly becoming a vital part of modern finance. The challenge is relatively simple – we alone cannot maintain pace with the continued flow of new data, nor the analytical systems required to make sense of it.
Alongside the rise of new information and an increased focus on compliance and efficiency comes an increased need for humans required to understand and manage that process from end to end. This process requires acquiring the necessary knowledge to deliver the datasets and the correct governance measures to facilitate and improve data quality. Under the GDPR, the requirement for explainability in these processes has become mandatory. Finance businesses, especially those larger established organisations, contain significant sets of valuable data with the incentive to utilise this information. However, the sensitive side of this financial information, creating data standards, and an ethics framework must be the focus before developing these areas. An ethics leader, someone responsible for maintaining human benefit and ethics lies at the core of AI innovation, is a critical part of delivering growth and generating the value expected from AI automation.
Instead of assuming AI will deliver the best insights, it’s critical to understand how and why the results are as they seem. Focusing on this ethical approach and ensuring wider adoption is vital to the role of an ethics leader.
While AI can perform many tasks without human involvement, it is critical that those creating, operating and making the decisions completely understand any potential errors before AI replicates them. With the ethical and governance-based foundation, finance teams will potentially automate bad decisions faster. Training, testing and continuous monitoring and vital to success. Training data applied to AI systems must not include bias to ensure no bias is replicated at a later stage.
While ethics professionals must manage the best possible practice, ethical AI requires a holistic approach to data literacy and ethics. These go together when creating and implementing assured AI, capable of supporting and complimenting human capabilities. A recent study by Alteryx discovered that 42% of UK employees responsible for data work saw data ethics as irrelevant to their role. The reality is, however, that data integrity and transparent decision-making are critical for any success in AI-focused insight generation.