The world is becoming ever more digital, and banks are taking advantage of that to offer their customers even more efficient, tailor-made services. One new technical option that is getting a lot of attention in this regard is artificial intelligence (AI). What potential does AI have to offer the banking sector, and how will it benefit customers? Are there barriers to increased use of AI in banks and if so, how can we navigate them?
Diverse applications
There are so many different potential applications for AI in banks, from optimising business processes, to improving customer experiences, to helping customers make sensible decisions. According to industry reports, more than two-thirds of banks and insurance providers on the German market are already using AI applications. Obviously, AI has already found broad use within the financial sector, as those within the sector have understood its usefulness for improving processes and services.
Potential and challenges
Artificial intelligence has a great deal of potential for use in the financial sector, particularly in terms of fast-tracking innovation and increasing productivity. AI can design products and services faster, more efficiently, at a lower cost, or with a greater degree of innovation. However, even though these advantages seem very exciting, we must also pay attention to the risks associated with these tools. To do this, the sector will have to maintain transparency, explain use-cases to third parties in a way that is easy to understand, ensure the underlying data is reliable, and of course that no discrimination is taking place. In order to do this, regulatory bodies that understand and can monitor AI developments must be involved in the process.
And of course, the political perspective must be taken into account. The European Commission’s approach to AI is based on two core principles: excellence and trust. In addition to avoiding risks, it is important to ensure that the EU does not fall behind when it comes to developing AI, but instead becomes a leader in the field. That’s why the European Commission is attempting to improve conditions for the development of AI models in Europe by providing financial support and making it easier to access the data required. One specific supporting measure taken within the European Commission’s approach to AI allows SMEs to access a world-class supercomputer in order to better train their AI models. This is just one example of the European Commission’s plans. Another is that the European Commission is putting more focus on a concept known as sandboxing, designed to foster cooperation between authorities and the market.
Video: Our event “AI in banks: opportunities and challenges”
In the meantime, implementation of AI and big data solutions has already led to a multitude of changes within the banks, in particular regarding customer interfaces, operations, and risk management. In addition to their use as part of bank operations, AI tools are now also supporting software engineers in their work. AI suggests workable code, meaning that programmers can complete their work faster. It also helps them ensure they are using clear documentation. AI is also being used to generate content, in particular when it comes to processing large amounts of data, for example for development forecasts.
Weighing opportunities and risks
Given the broad spectrum of potential uses for AI, we must evaluate which processes can be automated without issue, which will work well when carried out by programs and people working together, and in which cases the work simply must be done by a person. Generative AI, in particular, should not be considered a means of simply replacing people completely. Instead, AI is a great support tool, allowing people to make better quality decisions faster. Doing so will maintain the quality of services while also providing relief for employees during busy times.
Generative AI is currently improving at a rapid pace, and banks are using it more and more for software development. This increases productivity, on the one hand, but it also places much larger demands on organisational processes within the IT department, as they have to adjust to shorter release cycles with more frequent updates and test automation. In light of this, global software companies are turning more and more to cloud solutions, and it is likely that on-premise implementation will eventually become a thing of the past. This presents a challenge to financial service providers who will thus be forced to move into the cloud, which may trap them in a cycle of dependence.
Increased use of AI and cloud services also has an effect on energy consumption and IT costs. In order to measure this effect, we recommend integrating monitoring IT costs and capacities into the day-to-day processes of developers.
And of course, increased use of cloud services also results in additional regulatory burdens, meaning that not every use-case for AI makes sense. It is particularly important to weigh these considerations against the backdrop of competitiveness.
Data protection and sovereignty
As already mentioned in the context of cloud services, use of AI could mean entering into serious dependencies with software companies. This is closely connected to the question of European digital sovereignty in light of the fact that the market is dominated by non-European tech companies. It will not be realistic or practical for the EU to achieve complete independence from foreign providers. Instead, European businesses must enter into partnerships with tech companies in order to maintain their access to and expertise in relevant technologies. It is important, however, to have a plan B in the form of a specific emergency operational plan in case of global turbulence.
Unimpeded, cross-border data transmission is an important prerequisite for cooperation with foreign partners, and a central component of mutual trade agreements. However, exchanging data with countries with very different data protection standards is often difficult. In spite of this, a blanket requirement to store data locally cannot be considered expedient. Instead, it is a good idea to train employees to handle new technology or put more focus on data protection and security standards for professional use of platforms that are accessed individually. Doing so helps to reduce legal risks.
We need appropriate regulations
Regulation is, in and of itself, not an obstacle for digitalisation within the industry. According to experts, this is true even for the European AI regulation, which can provide a reliable legal framework for the use of AI-based processes and services while also serving to promote innovation. However, there are still unanswered questions as to how the regulation will be applied within the financial sector. Here, it is particularly important to find a good balance between the regulation’s general principles and special requirements for financial markets.
The principles proposed in the AI regulation are designed to create market standards and allow for the flexibility to react to new developments. After all, regulation does not require creating a new law for each new development. Instead, it should provide a flexible framework that can also take any future developments into account.
Conclusion
AI has already altered the financial services industry, and will no doubt continue to do so. In addition to facilitating extremely dynamic developments for banking processes, AI has the potential to significantly enhance customer experiences. However, implementing AI is not a trivial process: security, technology and regulations must all be carefully considered. Despite these challenges, the financial sector is in the middle of an exciting time. After all, AI has the potential to fundamentally change how we handle banking.
The Association of German Banks will once again be discussing AI in an event taking place on 19 February 2024 from 5pm to 7pm. The event is called “AI in banks: how will new competitors and innovative services for clients change banking in the future?” You can find more information here.
