Technology: Artificial Intelligence

  • Intelligence = The ability to learn and solve problems (Webster’s Dictionary)
  • Artificial Intelligence (AI) = The study and design of intelligent agents, where an intelligent agent is a system that perceives its environment and takes actions that maximize its chances of success. (Russell, Norvig)

Virtual assistants that tell you tomorrow’s weather or that suggest a good Indian restaurant, recommendation systems that show you products you might like to buy based on your previous purchases and Facebook tagging your photos, Artificial Intelligence is already well-established in our daily lives and is so ubiquitous that we barely notice it.

In a dystopian vision of this process, these super intelligent machines would soon surpass humans in their ability to understand or control. This topic has long been covered by science fiction stories and some prominent industry leaders have enhanced those fears. For instance, according to a new report from the Global McKinsey Institute, 55% of the current jobs will be replaced by artificial intelligence and machine-learning technology.

However, advances in machine learning have also opened new opportunities for progress in critical areas such as health, education and energy. Artificial Intelligence has the potential to improve the lives of many by solving the world’s greatest challenges and inefficiencies.  (National Science and Technology Council)

The future of finance

Will Artificial Intelligence also shape the future of the financial sector? Is Artificial Intelligence already used nowadays? On October 2nd  2018, a conference on the future of Finance was held at the Solvay Business School. It was organized by the Solvay Digital Society, a club that wishes to address the digitalisation of the business world. Three speakers were present and shared some elements to help the students answer these questions.

Andrew Pease (Senior Director Analytics – Deloitte)

Andrew is an expert in balancing business expectations and technical possibilities for innovative analytics implementations. His specialties include making advanced analytics digestible for business and helping organizations to develop a competitive edge through more and better data-driven analytics.

Instead of using the term ‘Artificial intelligence’, Andrew Pease introduced the notion of ‘Augmented intelligence’. Why should we augment intelligence? As humans encounter cognitive biases in their decision-making process, machines and analytics have the potential to overcome those. The term ‘cognitive bias’ was coined by Tversky and Kahneman in 1972. The two psychologists showed different ways in which human judgments diverge from the ‘rational choice theory’. For instance, the ‘Clustering illusion’ is the tendency to see patterns in random events. A gambler will wrongly choose red after a string of red, believing the color will appear one more time.

Andrew suggested to use AI when humans are biased but to keep humans on top of it! He also highlighted the fact that a wise utilization of AI was essential. Indeed, when one uses an algorithm on a specific dataset, a choice is made. Hence, Ethics and morality are still part of the process. Moreover, humankind still possesses abilities and features that machines do not have. For example, creativity and empathy allow humans to continue augmenting tasks that are now automated by AI.

On the one hand, the augmented approach offered by machines has many benefits for the financial sector. Using data to see if behaviors make sense has improved Fraud Detection, Anti Money Laundering and Customer Lifetime Value. On the other hand, a cautious use of AI is important because algorithms can reinforce inequalities. In credit scoring (= the evaluation of a client’s loan default probability), the machine’s probabilistic assessment might consistently put aside single mothers, who have a higher propensity to default.

Matthieu Remy (Founder & CEO – Easyvest)

Matthieu oversees the general management of Easyvest, its product development, and its client happiness. Easyvest’s mission is to help all Belgians grow their hard-earned money in a smart, simple, personalized, low-cost, and educative way.

Matthieu Remy introduced the ways in which AI was used at Easyvest. The firm attempts to tackle the pension challenge in Belgium. In the future, it is expected that half of retirees will have to live below the poverty line. Consequently, investing in performant pension plans is essential for all citizens.

First of all, AI is used at Easyvest to predict the stock market’s performance. The company has gathered 60 years’ worth of stock market history and is using this big data to find the statistical distribution of financial baskets. If the curve has followed the same distribution for 60 years, it is likely to remain similar tomorrow.

Second of all, Easyvest wishes to offer all clients consistent advice. However, advisors have limited availability. Therefore, to reply to all clients’ requests the company must increase the number of advisors. But the more advisors, the less consistency. Once again, the company used AI to solve this problem. The advisors progressively transfer their knowledge and expertise to the machine which is made available to chat with the clients (chatbot). As the chatbot is increasingly exposed to advisors replying to clients’ questions, it replicates their answers when the context or question asked is similar.

Peter Jacobs (CIO Netherlands – ING)

Peter is Chief Information Officer and board member of ING Bank Netherlands. He is responsible for all application development and maintenance activities within the bank in the Netherland.

Fighting against fraud and protecting clients’ money has grown to be increasingly difficult for financial services institutions as transactions became global and crime became organized in the last few years. The first strategy adopted by ING to tackle crime was manual detection which quickly became unsustainable. It was followed by rule-based detection. Rules are made based on existing/previous types of crime. Once criminals innovate, there is a mismatch between old rules and new crimes. Therefore, rule-based detection has a limited scope.

Moreover, it is difficult for a bank to determine a client’s ‘normal behavior’ because people are not always consistent in their transaction patterns. In that regard, AI presents tremendous potential. Because of the availability of large customer data and the machine’s flawless memory of events AI can be used to develop reliable behavior patterns that efficiently flag irregularities for specific customers. But in the end, humans are the ones who must act on events that the machine flags, it is still up to them to fight fraud and cybercrime.


In conclusion, the effectiveness of businesses and banks can be increased by building their capacity to implement AI to carry out their missions more quickly, responsively, and efficiently. However, humans still have a major role to play in the decision-making process and must keep on calling on their creativity. Moreover, the three speakers emphasized the importance of young people learning to code.


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.