Óscar Barba
Co-founder & CTO of Coinscrap Finance
The role of fintechs in introducing explainable algorithms
As main banks’ suppliers, we understand that the opacity of Artificial Intelligence models can generate mistrust and concern among decision-makers in the sector. Risk managers need to comprehend how AI works and what factors influence its outcomes to assess and mitigate associated risks.
Explainability is crucial to ensuring that decisions based on AI models are fair, ethical, and reliable. Entities must be able to justify and explain how decisions are made and what information is used to reach them. This is particularly important when resolutions can have a significant impact on individual rights and financial security.
Brief history of explainable AI: concept and birth of a new study area
The term was first coined by the Defense Advanced Research Projects Agency (DARPA) in 2016. As explained on their website:
“New machine learning systems will have the ability to explain their justification, characterize their strengths and weaknesses, and convey an understanding of how they will behave in the future. The strategy to achieve this goal is to develop machine learning techniques that produce more explainable models. These models will be combined with cutting-edge human-computer interface techniques capable of translating models into explanation dialogs that are understandable and useful for the end user.”
Moreover, as established by the European GDPR regulation (EU, 2016):
“The existence of automated decision-making must contain meaningful information about the logic involved, as well as the significance and the envisaged consequences of this processing for those interested.”
This regulation protects the user’s right to receive meaningful information about the logic applied in automated decision-making. To achieve this, data scientists must design accurate models and be able to explain algorithm predictions in a straightforward manner.
Benefits of AI explainability in banking risk management
By adopting this technology, banks can drive innovation and technological advancement in the industry, achieving more effective and cost-efficient risk management. Let’s now explore the key benefits for financial institutions:
1. Enhanced understanding and trust in the process
Explainability of AI provides risk management teams with a deeper understanding of how decisions are made and what factors influence them. This allows for a clearer view of associated risks and informed decision-making to minimize them.
Additionally, explainability fosters trust within the organization, as well as among regulators and customers.
2. Increased financial inclusion
AI models can be affected by biases and discrimination, leading to unfair decisions. Explainable AI enables risk management teams to identify and address these biases more effectively.
By understanding which data and variables influence decisions, teams can adjust and improve models to ensure greater fairness and increase the number of users through the inclusion of individuals previously excluded for no reason.
3. Enhanced decision-making and risk management
AI explainability provides the risk department with a deeper insight into the factors influencing decisions and outcomes. This enables informed decision-making and overall improved management.
Understanding how different data and variables affect results allows teams to adjust strategies and policies to optimize performance.
4. Compliance with current regulations
Explainable AI is essential for meeting regulatory and legal requirements in the banking industry. Regulators increasingly demand transparency and explainability in the AI models used by banks.
By implementing these models, entities can demonstrate impeccable regulatory compliance, avoiding sanctions and fines.
5. Improved customer experience and increased satisfaction
The use of these techniques also boosts engagement among banking users. Understanding the decision-making process and knowing which factors influence it allows customers to have more information about the products and services they have contracted.
This creates trust and transparency in the bank-user relationship, enhancing satisfaction with the service and strengthening brand loyalty.
Also find out how financial products are being improved through the AI
Implementation of AI explainability in the banking sector
After listing these advantages, are you considering introducing the explainable artificial intelligence (XAI) process in your institution?
If the answer is yes, here are some key steps.
Firstly, it is crucial to establish an AI governance framework that defines standards and guidelines for the development, implementation, and monitoring of AI models in banking risk management. This framework should include clear policies and procedures to ensure transparency and explainability of AI models.
Additionally, it is essential to consider that various techniques and tools are available to enhance the explainability of AI models in banking risk management. Some of these techniques include the use of Shapley values and sensitivity analysis to understand the contribution of different variables to decisions made by AI models.
It is also important to train staff in the concepts and techniques of AI explainability. Risk management teams must understand how AI models work, what factors influence their decisions, and how to interpret results. This allows them to make informed decisions and effectively explain results to other stakeholders.
Let’s not forget that the implementation of explainable AI models in banking risk management is an ongoing process. Teams in charge must regularly monitor and evaluate the accuracy of AI models.
This allows them to identify areas for improvement and adjust models as needed to ensure effective management.
About the Autor
Óscar Barba is co-founder and CTO of Coinscrap Finance. He is an expert Scrum Manager with more than 6 years of experience in the collection and semantic analysis of data in the financial sector, classification of bank transactions, deep learning applied to stock market sentiment analysis systems and the measurement of the carbon footprint associated with transactional data.
With extensive experience in the banking and insurance sector, Óscar is finishing his PhD in Information Technology right now. He is an Engineer and Master in Computer Engineering from the University of Vigo and Master in Electronic Commerce from the University of Salamanca. In addition, Scrum Manager and Project Management Certificate from the CNTG, SOA Architecture and Web Services Certificate from the University of Salamanca. He recently obtained the ITIL Fundamentals certification, a recognition of good practices in IT service management.
