Artificial intelligence – an assistant, rival or threat?
Many people still appear to believe that artificial intelligence (AI) is the technology of the future. However, it should be acknowledged that the future is here today, and that we are surrounded by a lot of AI-based solutions offering benefits to both consumers and businesses.
It is the financial sector that traditionally pilots and breathes life into many new technologies, including AI. Financial service providers, supervisors and central banks are currently busy dealing with the rapid development of AI-based solutions to enhance the improvement of financial services.
AI is a technology enabling a computer to think like humans. AI means that a machine can perform tasks and activities normally considered characteristic of the human mind, e.g. compilation of information, problem solving and experience-based learning. AI is not the same thing as a robot. It could be described as a computer located inside a robot and controlling it. For example, a robot would be a driverless car or a mechanical hand that can draw, while AI is the algorithm that decides on how to drive the car or what the mechanical hand will draw. Moreover, at the same time, AI is constantly improving its skills.
Primarily, we talk about two basic AI categories, but it is possible to identify even three of them. The first basic AI category is weak AI or machine learning specialising in one specific area, e.g. all it can do is play chess. Machine learning includes algorithms programmed to learn from the data obtained from the external environment, e.g. during a chess game, following each move made by the opponent, the machine has to select further game development scenarios aimed at winning the game. Currently, the most studied AI category is machine learning, and solutions based on it are all around us. The second category is the strong or general AI, which can perform the same tasks as humans, and it is smart enough to be recognised as universal AI rather than intended for one purpose only. Researchers acknowledge that the objective to be pursued was the attainment of this particular level, but it has not been reached yet. The last category – artificial superintelligence or the machine surpassing intellectual abilities of humans remains even further from reality today. This could be achieved by a machine which is constantly developing a better version of itself, outpacing humanity even in directions where it seems impossible at the moment, i.e. conceptual thinking, scientific innovation, empathy and social skills.
AI will play an increasingly important role in banking, which is currently facing major changes driven by the rapid development of technologies, changes in customers’ habits, the desire to receive services in the digital environment rather than in person and by new financial services providers that are more flexible, use new systems and challenge the old banks.
AI has a huge potential to change banking business models and develop new generation banks.
For example, AI can help to switch from the traditional revenues of the banking business provided by the traditional financial services, including payment services, to the supply of data-based products to individuals and businesses. These products would enable a more efficient customer service and generate new sources of income.
It is safe to say that AI will play a significantly greater role in our daily lives in the future. Already today AI helps to deliver individually targeted advertising to users of the social network Facebook and offers content that satisfies our potential interests on the streaming platform Netflix. AI also creates travel routes on Google Maps, helps floor robots (vacuum cleaners) to clean floors in an effective manner and lives in almost every smartphone in the form of assistants Siri, Bixby, Cortana or others.
The use of personal virtual assistants is rapidly developing in the West. Currently, the market is dominated by Amazon Echo and Alexa, but overall the supply is quite broad. Today, such assistants can provide useful information, e.g. weather forecasts, and order goods and services, e.g. food delivery. They can also be used to control smart home appliances, from TV and audio systems to regulation of heating systems. In the future, AI-controlled virtual assistants will be able to identify our needs and meet them without instructions.
Financial services providers and public administration are currently actively introducing other types of assistants, such as chat robots or chatbots, to help customers to find answers to topical questions. These virtual assistants are able to provide answers to frequently asked questions orally or in writing 24/7. However, opinions on this innovation are divided: some believe that a bank or provider of other services can thereby be better reached, others consider that an organisation tries to economise human resources without the consideration of whether chat robots really provide better customer experience in communication with a bank.
Another significant advantage of AI is its ability to process large amounts of data without any difficulties.
Unlike humans, whose decisions are also significantly affected by emotional and intuitive processes, AI is a technology fully based on experience and data, i.e. the higher the quantity and quality of data, the better the results produced by AI. Today, when the daily volume of data reaches its historical highs, i.e. in 2025, the volume of new information generated on a single day will climb to 463 exabytes (10006 bytes), the ability to process and use large amounts of information efficiently, including its combination with context, is very important. In the age of data and information, AI contributes to taking more informed and, accordingly, higher quality decisions in real time.
It can therefore be concluded that bank customers could benefit more from chat robots in the coming years when their development reaches the level necessary for assessing a customer's questions and requests concerning the bank products he uses as well as a customer's data and habits. As a result, AI will be able to offer better personalised solutions and products to the customer. AI will enable the bank to provide services in a smooth and accurate manner without requesting additional information from the customer, but instead using the analysis of his behaviour, habits and identity. Consequently, customer experience will be more advanced and their satisfaction – higher. Moreover, when the AI process is established, the need for human intervention in the AI decision-making process will be reduced to a minimum.
At the same time, AI can currently contribute significantly to development of the so-called regulatory technologies or RegTech in the financial sector through substantial optimisation of the costs associated with regulatory compliance and supervision.
In terms of robotics and AI topics, many people respond with suspicion and even fear: will mass adoption of robotics technologies combined with AI significantly reduce the need for people to work and, if that is the case, what will happen to people?
Researchers have forecast that 800 million jobs might be lost due to automation worldwide by 2030. Others believe that even half of jobs in the United States of America might be threatened due to automation (which is the subject of our colleagues' article Can we compete with artificial intelligence on the labour market).
The Augmented Human Enterprise, a study by the University of London, acknowledges that organisations augmented by automation technologies are 33% more likely to be human friendly workplaces, in which employees are 31% more productive. According to the study, the primary reason why automation makes people happier at work is the fact that "it frees them from repetitive tasks to concentrate on creative and strategic work – the bits of their job they enjoy most". According to Chris Brauer, Director of Innovation in the Institute of Management Studies of the University of London: "The future autonomy of digital technologies promises an evolutionary leap in our ability to grow as human beings."
The terms AI and automation are often used as interchangeable. However, automation may and may not be based on AI. Automation has one objective: to enable machines to perform repeated, monotonous tasks or, as some people say, "take a robot out of the human". It allows people to focus on more important, creative tasks requiring special manual work or personal evaluation. This will result in a more efficient, cost-effective business based on productive labour. An obedient robot never gets sick and works without holidays. Therefore, many companies are actively following the path to automation.
In turn, AI, as mentioned above, is intended to imitate human behaviour and intelligence.
The major prejudices in terms of AI stem from the fear that AI, which would be superior to humans, could solve many problems on our planet by destroying their cause – humans. There are also more realistic threats. i.e. future AI development might involve machine programming to create something horrible, such as fully autonomous weapons advanced enough to take final decisions without human involvement and used against people. However, for the time being, more attention is being paid to the risks, problems and disadvantages associated with the practical use of AI, which should offer benefits to humans rather than cause harm.
Overall, these risks could be described as legal by nature, e.g. the right to be adequately assessed on the technology side, AI transparency and explainability of the decisions taken. AI capacity is considerably higher than that of humans to enable it to detect hidden and delicate relationships that could be difficult for humans to understand. The result of AI work depends directly on the quality of the data used, which may result in situations where AI does not reduce the human-specific bias but, on the contrary, intensifies it. For example, if, historically, a company has received more male applications for the vacant post, it is highly probable that AI algorithms will produce a discriminatory assessment of women. There is a risk that such discrimination will develop even if conditions to prevent discrimination are entered in the algorithm.
The existing banking developments demonstrate that an AI-based system can take effective decisions.
Thus, it would only be logical for AI models to ensure such an opportunity provided that appropriate and fully-fledged testing has been carried out and control mechanisms have been established. Ultimately, the machine cannot be held liable for its operation, and despite the significant benefits derived from technology, we have to deal with it in a responsible and wise manner.
The next articles will focus on the use of AI in the financial field and how digital transformation changes the financial sphere and affects users of financial services.