Opportunities, Risks, and Challenges for the Financial Sector through Artificial Intelligence

Op-ed by Dr. Andreas Dombret and Dr. Rainer Glaser (published originally in German by 23rd of May 2024 at Der Bank Blog online at https://www.der-bank-blog.de/der-iphone-moment-finanzwelt/technologie/37708868/ )

Artificial Intelligence (AI) and particularly generative AI (genAI) have been on everyone's lips, especially since chatGPT. Opinions regarding the sustainable and structural impacts for the financial sector range from pessimistic skepticism to visionary euphoria or the expectation of a fundamental change in the working world, almost comparable to entering a new information age. The authors illustrate why reality might lie somewhere in between.

Opportunities, applications, requirements, and risks through AI are largely not yet fully understood - there is often a lack of understanding (and simultaneously no real awareness) in many places. We are still in an early stage of actually harnessing the potential of genAI.

Generative Artificial Intelligence: a hype or a fundamental change in the financial world? The fact that the entire spectrum of opinions is represented regarding this question is also because there is still much half-knowledge or ignorance shaping discussions and views. This also implies that currently, the potential and innovative possibilities of AI - generative AI being one aspect but by no means the only one - are far from being fully exploited or recognized. Of course, there are many "experiments" and "proofs-of-concept" that seldom make their way into operationalization, and at the same time, the range of "use cases" is limited and quite repetitive when comparing the ideas of different institutes. Genuine and sustainable innovation has not yet truly happened.

Therefore, the realization is slowly dawning that (gen) AI is about more than just triggering responses from a Large Language Machine or using genAI to pre-write texts or navigate through long documents oneself. At the same time, the wildest ideas and approaches circulate regarding what genAI will take over in the future, including the forecast that a large portion of current jobs will disappear - but here too, with little realistic, short-term prospects of realization, as the fundamental aspects are not internalized, thus the implementation ambitions are not achieved, or setbacks and local drastic misbehavior, possibly with material negative implications for institutes, seem programmed.

AI and genAI are powerful techniques - they will change the financial sector, both from the industry side and from the regulatory authorities. However, all of this will happen more as an evolution and less as a revolution.

Sustainable and secure utilization of the new possibilities requires a new skills profile that still needs to evolve across the hierarchy levels in institutes: expertise in AI or Data Science, combined with a deep understanding of business behavior and the associated value chain, supplemented by implementation and (model) risk competency. In our view, this is essential for the actual harnessing of the new techniques.

Only in this combination can the above obstacles be overcome, and genAI actually be tangibly implemented in practice. The powerful possibilities that genAI offers will gradually be realized; however, against the backdrop of the tasks outlined above, a "Big Bang" remains unlikely, as some predict for the financial world.

In our assessment, it is quite correct that the financial industry is particularly predisposed to derive material benefit from the new techniques. The focus on data, quantitative information, and other objective data sources, sometimes complex processes requiring a lot of expertise and effort, the necessity to efficiently manage customer information with a focus on customer benefits and efficiency or competitive pressure - all of this provides a suitable playing field for the broader use of genAI.

An excellent opportunity and at the same time a material challenge for banks and financial companies: long-standing challenges can be efficiently addressed. However, it is about much more than just solving age-old problems.

Therefore, genAI offers an excellent opportunity for players in financial markets to leverage their broad potential for improvement. We also expect that the wheat will be separated from the chaff: those who early on set the right levers and define and articulate a clear, realistic vision regarding the use and their own positioning in the genAI context will rapidly evolve and set themselves apart from the hesitators. This will not only result in better customer relationships and therefore better business success but will also be a significant competitive factor in the fight for young talent in the future. A successful positioning in the latter will further widen the gap with other players and simultaneously create new opportunities and innovation in service for customers.

We do not believe in the oft-repeated "work like a tech company" premise for financial institutions. Rather, we believe that the service philosophy of the financial sector will remain intact. However, the working model and culture will need to change to sustainably strengthen and further develop the central part of the mandate - customer service.

Even though it may seem surprising at first glance that we address the topic of culture in this context, in our opinion, it is nevertheless important and understandable: the seamless integration of the new AI capabilities into business models requires the development of a new mindset and a more interdisciplinary model of collaboration. Where we can already observe it today, we are confirmed that this is a significant and correct change: the "technicians" and the business experts will collaborate much more closely in the future and learn much more from each other than is the case today.

The new opportunities and chances do not come without a price. Like with any technology, handling must be learned, and risks must be identified and managed. An additional requirement for the competency profile of institutes, their employees, and the regulatory authorities naturally accompanies this.

The aforementioned tension between fear and euphoria is probably precisely the best starting point: finding the right balance between both extremes is crucial because only in this way is a targeted and sustainable approach to realizing AI procedures feasible. This sounds simple, but it is not always the case in practice: both ends of the spectrum must be much better understood and structured. On the one hand, the spectrum of possibilities is far from being fully recognized; on the other hand, the associated risks must be understood and identified more precisely. Both dimensions must, therefore, evolve hand in hand - including the roles and requirements of 2nd line functions and standard setters as well as regulatory authorities.

This learning process will certainly not take place without individual and possibly even material failures. From these, lessons will be learned, and they will shape best practice standards. Against the backdrop of the evolutionary systematic and the diversified roles or ambitions that the various participants in the financial sector will take on, it is not to be expected that artificial intelligence, as expected by some, will trigger systemic events in the financial sector or even cause the next financial crisis.

AI-driven change will come and has already begun, and it will fundamentally change the financial sector, setbacks and negative events cannot be avoided in this process, but overall, in our firm conviction, sustainable benefits for the financial sector and consumers will emerge. But: all of this does not happen on its own and does not come without a price.

AI will positively advance the financial world as a whole, and this advancement will accelerate further - not only because the understanding of harnessing will rapidly increase but also because the technology itself will take further significant steps forward. The authors of this guest post stand on the optimistic-realistic side in this regard: it is from our perspective correct and important to utilize the new possibilities through AI broadly and with a focus on benefits, and with the right approach, this offers significant and sustainable advantages. However, the right approach must be defined individually in each case, the path to it must be relearned, corresponding standards - both from regulatory authorities and from the institutes themselves - must be redefined, and real agility and a completely new way of thinking will be indispensable.

Dr. Andreas Dombret was a member of the Executive Board of the Deutsche Bundesbank and is currently, among other roles, a Global Senior Advisor at Oliver Wyman , Dr. Rainer Glaser is a Partner at Oliver Wyman in Frankfurt.