While there’s no doubt that AI adoption is on the rise in banking, there are still some challenges the industry needs to overcome. Some critics have argued that AI in banking comes at a cost – the disintermediation of banks from their customers and loss of the ‘human touch’. One of the biggest barriers to AI in banking has also been the lack of transparency around how AI decisions are made.
In this regard, Apple recently made headlines for allegations that its Apple Card uses biased algorithms to set credit limits and as a result discriminates against women who apply for the card. It was even claimed that men who apply for the card receive between 10-20 times higher credit limits than their wives. The outcry led to reputational damage and an investigation by the New York State Department of Financial Services.
Outside of the financial industry, others have also fallen on the wrong side of “AI bias”. Amazon, for example, tried building an artificial intelligence tool to help with recruiting, only to later find it discriminated against women because it had combed through male-dominated CVs to gather its data.
Perhaps most worrying, however, is the potential for these flawed, opaque AI systems to ‘teach’ themselves, reinforcing bias as their decision making develops. This problem is exacerbated by the investment in opaque ‘black box’ AI systems, which cannot communicate how decisions have been made to the operator, regulator or customer.
Since these ‘black box’ AI systems rely on data, they can learn from interactions and rapidly accelerate poor decision making if the data they are fed is corrupt.
The only solution to this is ‘white box’ explainable AI (XAI) systems which explain in simple language how the software operates, how decisions have been made and that are able to answer follow up questions in order to maximise customer financial wellbeing.
Transparency is key. By explaining how and why decisions are made, XAI helps consumers and companies understand what they need to do to get a different outcome. This could involve turning a rejected mortgage application into an acceptance, for example.
The technology helps consumers take appropriate action, while also opening new business avenues for banks and other institutions who have the insights they need to offer more suitable products.
The potential that XAI has to improve customer experience in the banking industry is huge. Today, crucial decisions are already made by AI on loans, wealth management, and even criminal risk assessments. Other key applications of AI include robo-advisory, intelligent pricing, product recommendations and debt-collection.
Easy to see that XAI is set to play a key role in banking
When we consider the real value that AI brings to the table, it’s not hard to see why its use is becoming so widespread in the global financial industry. It has the ability to sort through vast amounts of data in real-time and process both structured data (such as a filled out forms) and unstructured data (such as a voice messages). And let’s not forget that it can do this much faster than a human and is less prone to error.
Regulators see this value and recognise that in this era of open banking and PSD2, there is a real need to provide a framework in which the vast amounts of data being shared can be used to provide customer-centric solutions through intelligent AI-driven algorithms. Because of this, we are likely to see stronger regulations in the future that ensure AI algorithms do not apply bias and remain transparent.
In this future scenario, it’s not hard to see why XAI technologies are set to play a key role.
Looking towards 2020 and beyond, digital banking channels and AI technologies will increasingly become the main way that customers interact with banks. If not handled correctly, customer disintermediation could plague the banking industry. However, by adopting customer-centric technologies like XAI, the customer intimacy that’s played such an important role throughout the history of banking can not only be regained but taken to new heights.