Tech, a new challenge for sustainable finance

When talking about ? sustainability ?, environmental and social commitments of a company is what comes first to one’s mind. While this notion represents without a doubt a major point, there are in my opinion two other layers of sustainability that should not be neglected by the banking sector. Firstly, financial and regulatory sustainability, since this is the foundation of a sustainable financial environment and the potential systemic risks that the financial sector represents for our economies as illustrated by the financial crises between 2008 and 2010. And finally, and that’s the topic I would like to focus on, the new frontiers of sustainability and notably technological sustainability, a subject which in my opinion must become a major part of the discussion around sustainability and positive transformation.   

Data, robotics, AI are the words present in all minds. These new technologies bring new risks that banks are to face and accordingly, the underlying social impacts.

Let’s start with data, a perfect buzzword. A lot of issues could be put behind it. Numbers are striking: 90% of global data has been created in the past two years and the amount of data we generate every day is estimated at 2.6 billion terabytes [1]. In this context, it is mandatory for financial players to address the issues around data creation and protection. 

With no surprise, regulations (ex. GDPR) in this field are becoming stricter. Fortunately, banks and insurance companies benefit from the reputation of being a trustworthy partner. The challenge here is to deliver on this promise and keep the advantage while the temptation to monetize this “asset” through selling it or using for proper marketing purposes is huge. 

The second domain that transforms all industries is robotics and its impact on the job market. Recent studies estimated that by 2027, from 23% to 50% of the total workforce in the financial sector will be affected by job cut (replaced by AI and robots) [2]. But when we have a look at the example of German individual manufacturing workers, studies[3] tell us that there is no evidence that robots cause total job losses, but they do affect the composition of aggregate employment. Concretely speaking, every robot destroys two manufacturing jobs but creates additional jobs in the service sector thus fully offsetting this loss.

Actually, no matter what the impact will be in terms of numbers, we are clearly in a Schumpeter's process of “Creative destruction” [4]. Robotics and AI will modify the equilibrium between the type of jobs their location and the skills required and have a disruptive effect that must be carefully handled in the affected countries.

In this context, banks themselves should ensure to anticipate the impact of these technologies on their employees and – consequently – accompany their workforce in the transition through investments in training and career management.

Artificial intelligence is probably the most difficult topic to address due to its complexity and its ethical implications.

Ethics is not new for banks but with AI we are moving to another level where we must anticipate potential risks and define the mechanisms to ensure control and accountability. Here, I see two major ethical issues: algorithm biases and transparency of the models.

Algorithm bias or AI bias is a phenomenon that occurs when an algorithm produces results that are systematically prejudiced due to erroneous assumptions in the machine learning process. Thus, the growing use of algorithms opens opportunities but also the possibility for unfair bias. Let me give you some examples. In 2014, Amazon developed a recruiting tool for identifying software engineers it might want to hire; the issue was that the algorithm was configured using the CVs of the existing pool of engineers who were predominantly male, thus the system swiftly began discriminating against women, and the company abandoned it in 2017.

More recently, Apple launched with Goldman Sachs an Apple credit card accused of being sexist. Indeed, for a married couple who files joint tax returns and live in a community-property state, the Apple’s black box algorithm decided that the husband deserved 20x the credit limit compared to his wife.

Algorithms can have built-in biases because they are created by individuals who have conscious or unconscious preferences or fed by samples of data incorporating that kind of bias. And they may go undiscovered until the algorithms are used. That is why the implementation of the right safeguards while developing the models is crucial.

Besides build-in biases, the other ethical concern is about the transparency and the “explainability” of these models. Because the direction is clear that these models are and will be more and more used for making significant decisions, be it HR recruitment, a credit granting and may be tomorrow even for a judgement. It becomes then essential to know the considered features and their importance in a decision. We need to open the black box, understand the decision-making procedure, and evaluate the inherent assumptions made by the models. Above the ethical part of the issue, we are also pushed by a reglementary one, as the EU General Data Protection Regulation introduces a right for individuals to obtain “meaningful information of the logic involved” in automated decision-making with “legal or similarly relevant effects”.

Technological sustainability concepts still represent today more questions than answers.

Indeed, we are actively working on regulations while climate issues are rightly becoming more and more present in the sectoral debate. However, we should now look further: is the profession well organized around the technological sustainability and the ethical implications?

The solutions to this almost philosophical question require collective thinking, actions and framework. It must be a shared effort for future bringing together public and private sectors as well as consumer groups.

Personally, I am convinced by the winning combination of Human and Digital.

While we consider all the risks, we should keep in mind that, globally, this technological progress brings better lives for everyone. It has vast potential, and its responsible implementation depends on us.

[1] https://www.forbes.com/sites/bernardmarr/2018/10/01/what-is-deep-learning-ai-a-simple-guide-with-8-practical-examples/#25d4ab3b8d4b

[2] According to BCG studies & “Finance leaders say AI will soon replace half of banking sector’s jobs” https://www.personneltoday.com/hr/finance-leaders-say-ai-will-soon-replace-half-of-banking-sectors-jobs/.

[3] German Robots – The Impact of Industrial Robots on Workers, “IAB Discussion Paper” published by the research institute of the German Federal Employment Agency.

[4] The "gale of creative destruction" describes the "process of industrial mutation that incessantly revolutionizes the economic structure from within, incessantly destroying the old one, incessantly creating a new one".