BBVA and the University of Navarra have partnered to conduct a three-year applied research project into responsible AI.

Titled Fair Learning, the project aims to investigate the main ethical, technical, and regulatory challenges posed by the growing automation of decisions that impact society.

BBVA said a multidisciplinary team of 20 experts including data scientists, academics from fields of law, economics and philosophy, will take part in the project.

One of Fair Learning’s main avenues of research is how to mitigate the presence of bias in data and BBVA said the team will utilise advanced mathematical and statistical methods to help prevent models from learning discriminatory or biased patterns while being trained and monitor the results and outputs of trained models to ensure that they do not infringe the interests of any population group.

The team will also integrate philosophical frameworks into the project, such as human-centred AI and virtue ethics, which guide technological development based on moral criteria and social responsibility.

BBVA said the aim is to build systems that respect privacy, promote fairness and reinforce human autonomy without replacing it. With decision-making processes becoming increasingly automated, including loan approvals, recruitment or access to services, BBVA said banks must ensure that artificial intelligence systems are both fair and transparent.

In the regulatory realm, BBVA said the Fair Learning project will analyse the existing legal framework, which is built around the new European Union AI Regulation including the EU AI Act, to define best practices and propose recommendations to ensure ethical and legally aligned implementation of AI systems.

“This is a genuinely groundbreaking project in the financial sector and a further show of BBVA’s commitment to advancing its technological transformation without losing focus on people,” said Josep Amorós, project coordinator and senior manager of analytics transformation at BBVA. “AI is already a central pillar of the bank’s transformation and will only gain in prominence over the coming years. Therefore, we must ensure its development adheres to principles of fairness, responsibility and transparency.”

Fair Learning is part of a wider ongoing partnership between BBVA and the University of Navarra in the field of data which began in 2020 with an agreement to train the bank’s employees in data science, create a dedicated track for BBVA employees in the official Master’s in Big Data Science, and promote joint industrial PhDs between the bank and the university.

To date, more than 90 students have graduated from the master’s programme, 135 data scientists have been certified, and 12 PhD students are currently enrolled, bringing the number of BBVA professionals specialized as a result of the partnership to 239.

In November last year, BBVA announced it had extended its partnership with the University of Navarra to launch a training programme for over 150 managers to improve productivity through the use of generative AI (genAI).

The course has been created by the Spanish bank as well as the Data Science and Artificial Intelligence Institute (DATAI) at the University.

---