Artificial intelligence and banking risk management

Banks’ digital transition goes also through artificial intelligence. Credit institutions are using AI to reach three goals: first, to strengthen and increase their customer portfolio; second, to understand, foresee and protect themselves from risks associated with financial activities and external risks, such as Covid-19 pandemics; and third, to explore market opportunities and enhance profits, to become more robust and competitive.

To hit these three targets, the use of data, however big or comprehensive they are, is crucial. Beyond the traditional or alternative sources, data collection, management and storage are those steps of the process allowing banks to obtain quantitatively and qualitatively relevant data for the purposes of savings and credit activities.

The AI quickens and facilitates end-to-end operations on data, however raw or refined. This will be even more true in the future, as the weight and importance of information technology will exponentially grow thanks also to ultra-wide band connectivity, which is bound to revolutionize the banking sector as well as other fields of economy.

What is artificial intelligence

In general, artificial intelligence (AI) is a mix of technologies allowing computers to carry out a range of advanced functions, including the ability to identify, understand and translate the written and spoken language, to analyse data, express recommendations, and so forth.

We could say that artificial intelligence (AI) enables machines to think in an autonomous and more sophisticated way. For instance, thanks to AI computers can acquire data and extract strategic information which cannot be identified through conventional statistical analysis, using complex algorithms to formulate predictions and trends.

Artificial intelligence’s technology has created new opportunities for progress in some critical fields, such as health, environment, industrial automation and finance. In some cases, artificial intelligence can carry out operations more efficiently and methodically than humans.

4 types of artificial intelligence

Learning in AI can be ‘narrow’, ‘general’ and ‘super’. These categories show AI’s abilities as it grows, executing a set of strictly defined tasks (‘narrow’), the same thinking ability as humans (‘general’) or performances exceeding human capacities (‘super’). Thus, there are four main AI types, as defined by Arend Hintze, researcher and professor in integrative biology at US Michigan State University:

1. Reactive machines
2. Limited memory
3. Theory of mind
4. Self-awareness

Although the latter does not exist yet.

Artificial intelligence (AI) in the banking sector

Which are AI’s main applications in the banking sector? As it formulates models and foresees outcomes and trends, AI is essential for risk management. So much so that, as shown by a 2022 research of the Cambridge Centre for Alternative Finance and the World Economic Forum, most banks and financial services companies stated that they had implemented AI technology for both risk management (56%) and income generation through innovative products and business processes (52%).

Artificial intelligence allows credit institutions and FinTechs to better understand and mitigate more efficiently the risks associated with their activities. By using sets of big and complex data, data mining technologies and AIs may help banks to develop more accurate and precise risk models than those based on standard statistical analysis.

Through the assessment of a wider and more comprehensive range of data, for example, banks may gain rapidly and more easily strategic information which are useful to protect from losses and frauds.

AI’s benefits in the banking sector have been examined also in a Bank of Italy’s study focused on the use of artificial intelligence and machine learning techniques supporting the credit risk assessment conducted by Italian intermediaries. The Euro Banking Association (EBA) also intervened on the subject.

AI’s benefits for banks

According to an international survey conducted by the Canadian firm Open Text Corporation, eight out of ten banks are aware of the artificial intelligence’s potential advantages for their activities. And, in effect, several banks globally are planning to implement AI solutions.

Moreover, based on this study, there are three main channels through which banks can use artificial intelligence in their operational processes:

• front office
• middle office (to detect frauds and manage risks)
• back office (subscription)

In some cases, the use of artificial intelligence has already gained importance in banking transactions, such as chat-bot in front office and payment anti-fraud services in middle office, where AI is also a valuable ally in investment banking activities.

Let us not forget that today many banks use artificial intelligence to improve customer experience, ensuring 24/7 smooth and easy interactions with the bank itself.

Getting back to the security issue, Deutsche Bank’s model for countering financial crime has been paving the way since 2019. In fact, the artificial intelligence system ‘Black Forest’ analyses transactions and records suspect instances. For each capital movement, for example, several criteria are considered: amount, currency, the country towards which capital is directed and the type of transaction, whether the transaction occurs online or at the desk. This AI application has already allowed to discover many frauds and tax evasion occurrences, including one related to organized crime and money laundering.

Artificial intelligence and sustainability

AI is a precious tool for banks on the front line to support green transition as well as defend and safeguard the Earth.

The ability to quickly process large amounts of data makes artificial intelligence models useful and resourceful, even in a field like sustainability. Environment is one of the three pivotal elements of ESG, that is, an acronym which is becoming ever more relevant in current economy; to the point where, as of 2023, EU banks will have to report ‘green’ assets, calculated on the basis of new Community standards.

For instance, driven by EU regulations, credit institutions will have to indicate which funding, among those granted, will contribute to solar energy and wind power production and, as a consequence, could be classified as green. They will also need to measure and assess funding granted to medium-sized undertakings investing in production equipment or systems, which will potentially make them more climate change friendly. For such operations, banks need to collect large amounts of new data from their customers.

‘Until now, these data from customers had to be individually examined by consultants’, stated Murat Cavus, while developing new technologies to support Deutsche Bank’s sustainability efforts. In the future, AI application to machine learning will contribute greatly to the classification activities of the German banking giant. An algorithm will pre-select the information required for these processes – the so called autoclassification. ‘In this way, we will unburden our consultants from a large amount of work’, explained Cavus.

It is true that, in the end, a consultant will have to ultimately approve the suggestion developed by the algorithm. However, ‘autoclassification provides additional information which facilitate the final decision’, clarified Cavus. This is what Deutsche Bank imagine will be the future of AI in the institution: ‘We will increasingly use technologies to shorten standard processes, becoming in turn more environment friendly. This will allow us to better serve our customers and turn into a greener bank ourselves’.