The future age of AI

By John Morgan
Banking & Finance Financial Services Insurance Private Capital Superannuation

In brief

Written by Consultant John Morgan

We will be able to cease speculation on what is in or not in the Final Report of the Financial Services Royal Commission on Monday evening (or after whatever time it takes to read and digest!). As clarity emerges, we suggest one important matter that is worth considering – the impact of the future age of AI.

The Commission has looked at past events, and although some practices may be continuing, its recommendations for the future will be based on its consideration of what has happened. These events have largely involved traditional methods of product manufacture and communication.

The challenge for those responding to the Commission's recommendations is that they are unlikely to have very much regard to emerging practices and the risks associated with them. If we jump forward some years, we can imagine the major developments in the use of electronic systems for the creation, dealing in, marketing and distribution of financial products. This may involve the use of artificial intelligence, learning systems and complex algorithms. Think about the emerging connected autonomous vehicles or driverless cars which will operate by using their own systems together with signals from other vehicles, road infrastructure and the cloud. These onboard and external systems will also learn and change the behaviour of the vehicles. Won't financial services develop along similar lines with smart algorithms guiding even retail customers in their investing, insuring and borrowing activities?

Some of this future is a little way off, but in formulating regulation, supervision and enforcement after the Commission, we need to think about the behaviour, not just of financial advisers, bankers and other managers, but also how the regulatory system will respond to and hold accountable those who are the designers and operators of these systems. How does the holder of a financial services licence deal with these emerging electronic systems and ensure that they do not fail the requirements that a product is suitable for a customer, or that they are providing services 'efficiently, honestly and fairly'. The computer says no! – with apologies to Little Britain.

In the case of very sophisticated AI systems, the difficulty is that experts say that they do not understand how some of the emerging artificial intelligence systems actually work, in the sense that it would take many persons years to work through the millions of steps taken by the system in the second it takes to reach a decision or outcome.

Like the frog in water that is heating up as we respond to the likely changes, it is easy to see how we may not anticipate the need for regulatory, management and governance systems to be sufficiently flexible to absorb and respond to these and other emerging changes. We might bring about better behaviour that avoids misconduct and meets the expectations of today's customers by following the recommendations of the Commission but the more challenging task may be to mould the changes to also reflect and deal with the new AI age. And avoid the next Royal Commission.