In the dynamic environment of banking, technologies based on artificial intelligence and machine learning have become an integral part of digital transformation. They open up new opportunities, but also present challenges that require careful attention and a cautious approach – observes Waldemar Szczepański, Deputy Director of the IT Management Office at Bank Pocztowy, in charge of the bank’s development in the area of cloud technologies and artificial intelligence.
The use of artificial intelligence (AI) and machine learning in banking opens up numerous opportunities for the financial sector, but at the same time presents significant challenges. A responsible approach of financial institutions to the implementation of new technologies, taking into account both their potential and associated risks, becomes crucial. Focusing on the possibilities that the development of AI brings to financial markets, five most important aspects should be emphasized. Firstly – personalization and data analysis. The ability to process large amounts of data allows banks to create personalized transaction analysis, assess credit risk and detect fraud. The second aspect is the automation of routine actions, which leads to reduced operational costs and increased efficiency of internal processes. Customer service is another area where AI plays a significant role. Chatbots and voicebots allow the automation of customer service (self-service), which translates into increased service availability and improved customer experiences. Artificial intelligence can also provide online and offline assessment of customer conversation sentiment, allowing for better adjustment of provided services. Moreover, language models allow for the construction of personalized services and offers for customers, which are later used by banking advisors.
The use of artificial intelligence in the financial sector has many of the benefits already mentioned, but is also associated with threats that could affect the stability and security of the financial market. The most important issue to address is data security. Therefore, banks have teams responsible for data security, whose main task is to maintain the highest standards of quality and safety in order to eliminate risks associated with the use of AI, such as data leakage and violation of customers’ privacy, which would pose a significant threat not only to the banks but also to the customers themselves. Another potential problem for the industry is identity theft. AI technology brings numerous possibilities in, for example, voice or graphic manipulation, which is why effective protection mechanisms are required against potential fraudsters.
Financial institutions must take full responsibility for the data provided by artificial intelligence. Changes in the model can have a significant impact on the serviced processes – risk, credit decision-making, fraud analysis, etc. The threat and risk is not only associated with the possibilities brought by artificial intelligence, but also with potential neglect or delay in implementing AI in financial institutions. This action can result in the loss of benefits from using artificial intelligence, which has greater potential than standard improvements in internal processes.
It is crucial for financial institutions to take appropriate precautionary measures and monitor the risk associated with the use of AI. Effective supervision and control of AI-based systems are also necessary.
Robotization and the use of new technologies for reporting
Systems based on artificial intelligence can automate the process of preparing and sending reports to specific institutions. Thanks to advanced algorithms and data processing capabilities, they are able to effectively analyze information, generate reports, and even automatically send them to the appropriate recipients, thereby eliminating the need for manual report creation and dispatch, which significantly saves time and human resources. It should be stressed, however, that robots can support people, but they cannot replace them entirely. Our experience, including from the implementation of the Lady Robot project at Bank Pocztowy, shows that the key is to find a balance between automation and human knowledge and abilities. The challenge is also responsibility for the preparation and delivery of data, which at the end always belongs to the model owner. Responsibility can be split at several levels: responsibility for the quality and reliability of the model; responsibility for preparing the final contents; and responsibility for data sources and copyrights. Consequently, it is necessary to establish clear policies and procedures regarding responsibility for generated content, to ensure safety and compliance with the law.
