In June, one of the largest banking groups in the world – Lloyds Bank – announced it was joining the ranks of Google, KPMG and Uber with the creation of a new role… Group Head of Data Ethics. In short, a digital cartographer responsible for the bank’s moral and legal obligations for gathering, analysing, protecting, and utilising the vast quantities of data at their disposal.
The job posting specified the need to “promote, embed and commercialise Data & Analytics practice and culture across the business to enable a data-enabled organisation.” Along with driving the adoption of data ethics methodologies, acting for the “betterment of UK society”, and “using data (and data ethics) as an enabler to meet strategic objectives”.
If data is the fuel needed to drive modern business insights, then AI is the turbo engine that powers this process – delivering new solutions and insights far faster than previously possible. Conversely, however, using AI automation at this scale without a data ethics leader in such a highly regulated environment is the equivalent to starting a road trip with no map, no sense of direction, nor an ability to adjust course to get back on track.
With a consistent and all-encompassing requirement for audits, compliance checks, and explainable processes, consistently applied standards in the financial sector have been a historic safeguard of financial success – not only in regulatory fine avoidance, but also to avoid lawsuits due to data errors. So, what happens when the speed of data creation outpaces the ability to process it effectively?
The meteoric growth of data generation and the potential of automated insights
To understand why data ethics is so important, we must first consider the role that data plays within the finance industry. In combination with the complexity and volume of data being generated in business today, this influx of data quickly outpaced the original infrastructure and processes needed to effectively analyse it.
Where compliance and regulatory processes were once completed manually through dozens of static spreadsheets, today that process is becoming more streamlined through the use of advanced analytics technologies that are much more approachable and accessible. With every organisation essentially becoming a data factory, analytics automation is the way to refine the billions of rows and thousands of columns of data swiftly into actionable insights.
Arguably, the two most significant tools in use within the finance sector today are business process automation and automated insight generation from datasets. Process automation is, namely, the ability to take a specific task – such as pulling data from its source and combining it with data from other areas – and automatically build reports. This has far-reaching benefits for compliance and regulatory requirements and can cut report generation time down by 99% with almost no errors. This is due to repeatable, transparent and easily verifiable steps completed in the exact same way each time.
The second significant tool is the use of AI for decision-making. AI can be used to detect patterns within datasets to identify fraud or money laundering, or even to highlight specific interdependent factors which may impact an applicant’s ability to repay their mortgage. The efficiency and speed benefits that automation brings means that this technology is fast becoming a staple of the modern finance function.
The challenge faced is contextually simple. The human brain alone simply cannot keep up with the unending flow of data, nor the analytics technologies required to make sense of it.
Setting the foundations of ethical decision making through AI
While AI promises to provide significant benefit, the inverse can also be true without the right foundational attributes such as analytics upskilling or data literacy. Alongside the growth of data created, and an increased onus on compliance and efficiency, also comes an increased need for the humans involved to understand and manage that process from end to end. This requires not only the right business domain knowledge to deliver non-biased datasets, but also the correct governance foundations to facilitate and amplify good quality data work. Notably, under the GDPR regulation, the need for explainability in these processes has become mandatory – particularly where PII is concerned.
Banking and financial institutions – particularly those with long histories and a large user base – are sitting on a mountain of valuable data with a huge business incentive to make use of it. But given the sensitive nature of this financial data, setting data standards, governance and ethics framework guidelines must be a priority before developing these functions. An ethics leader – one responsible for ensuring human benefit and ethics are at the centre of AI innovation – is the essential ingredient for delivering growth and creating the value promised from AI automation.
Rather than believing AI will simply deliver the correct autonomous insights, it’s imperative to fully understand how and why it arrived at the answer it did. Delivering this ethical foundation and ensuring widespread adoption is core to the role of any ethics leader today.
AI models, explainable outcomes, and the human intelligence factor.
Although AI can be trained to perform many tasks without human interaction, it is essential those designing, operating, and making decisions from the outputs fully understand any possible defects before AI amplifies them. Without that ethical and governance-based foundation, finance teams will only be automating bad decisions faster. Training, testing, constant monitoring, and continuously enforced standards are integral to success here. Training data used to feed these AI systems must be free of bias to ensure said bias is not simply replicated down the line.
While ethics leaders shoulder the responsibility for spearheading best practice, ethical AI requires a company-wide approach to data literacy and ethics. These attributes go hand in hand when developing and deploying trustworthy AI capable of augmenting and complimenting human capabilities. Alteryx commissioned research into the state of data literacy in the UK found that, shockingly, 42% of employees responsible for data work saw data ethics as “irrelevant” to their role – casting a shadow over future AI-based projects. But data integrity and fair decision-making are cornerstones of any successful AI-based insight generation.
The future of ethical AI stands at our doorstep, and wide-reaching data literacy underpin its progress. Statistics from GlobalData show that more than half the data ethics roles advertised in May and June were with financial organisations. Considering the huge upside to getting AI automation right, the only surprise at this point is that data ethics is not more prevalent in such a highly regulated sector.