The fast development of Artificial Intelligence is a source of progress but also triggers many questions. How would you describe it?
A generally accepted definition assimilates Artificial Intelligence to “the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages”.
At its simplest form, AI combines Computer Science and Data to enable problem-solving. But AI is associated more and more frequently to learning. It is through learning, and experience, that an intelligent system capable of executing a task can improve its performance and will be able to perform new tasks and develop new skills. Thus, automated machine learning and deep learning are the two subfields that are frequently mentioned in conjunction with artificial intelligence.
« We must ensure that AI is deployed in our organisation in accordance with our values and code of conduct. »
With the explosion of data produced by our society and the exponential increase in computing power, AI is nowadays providing cutting-edge technology that serves innovation in virtually all the economic sectors.
Like many actors, BNP Paribas is embracing AI to improve its operational efficiency and the way it is serving its customers. However, as with any pervasive technology, we must ensure that AI is deployed in our organisation in accordance with our values and code of conduct.
Concretely, this means that we strive to:
- focus on ethical dimensions and ensure that AI is not reproducing stereotypes or social or cultural discriminations
- ensure that algorithms do not infringe on fundamental human rights – from privacy and data confidentiality to accessing essential services
- create the conditions for trust
- promote an AI that is human centric
In a nutshell, we want to have, through and thanks to AI, a positive impact on all our stakeholders.
How is AI applied within BNP Paribas?
At BNP Paribas we are using AI in a wide variety of domains. We are focusing on projects that aim at improving customer knowledge, experience, and services. For example, chatbots implemented in our networks use AI to understand our customer issues faster and provide more efficient answers.
Recently Cetelem Brazil implemented a very effective chatbot within the social messaging platform WhatsApp. Its aim is to ease customer interaction by providing fast and real-time responses, guiding them to our products and services and enabling 24-hour transactions.
Robo-advisors, or virtual advisors, can also be developed with AI. A good example of this is the acquisition by BNP Paribas Asset Management of a majority stake in the fintech company Gambit, thus allowing us to offer the clients of our retail and private banking networks investment advice based on the power of algorithms.
We also have projects that enhance BNP Paribas’ operational workflows, such as process automation or decision support.
Finally, we use AI to improve our risk management and compliance practices, like fighting fraud or money laundering. Today, AI is a widely spread and maturing practice with dedicated teams and competencies throughout most our business lines and functions.
Why is it crucial to ensure an ethical development of AI?
AI is a powerful enabler of innovation and efficiency provided its use is grounded in sound values and ethics. For me, ethics is knowing the difference between what you have the right to do and what is right to do. When it comes to AI and by capillarity to data -as AI cannot happen without data-, it is what makes the difference between a compliance approach versus developing a true culture supporting the responsible and sustainable use of AI and data.
However, it is fair to say that humans are hardwired for bias, be they conscious or unconscious. Those cognitive biases are deeply rooted into many aspects of our economic and social environment: workplace, healthcare, scholarship, law enforcement, marketing - among others. It is not surprising, then, that bias also infiltrates every byte of the data our society produces and consequently, the algorithms trained on this data. This risk is unfortunately real and too many incidents have sadly shown that AI models, of different types, can reflect racial or gender biases.
« AI is a powerful enabler of innovation and efficiency provided its use is grounded in sound values and ethics. »
For this reason, it is of tremendous importance to establish the appropriate safeguards to avoid their materialisation. This is also why in September 2020 - as part of our strong commitment to ensure a responsible use of AI - we signed the Women's Forum pledge for an inclusive AI.
And concretely, how is BNP Paribas dealing with those stakes?
I am strongly convinced that ethics come with awareness, and it can’t be the sole responsibility of a dedicated team. It should involve every one of us. This comes with the culture of understanding what data is and how to use data.
In practice, for BNP Paribas using data implies:
- transparency in the way we collect and process data.
- developing privacy-by-design and privacy-enhancing products and infrastructures.
- giving individuals power over their data and through their data
- ensuring that social, gender or other biases that exist in our societies are not adversely affecting people, in particular, those that are vulnerable
In simple words, it is about treating someone else’s personal information as you wish your own, or your childrens’ information would be treated.
With the Women’s Forum, pledge, we committed to delivering a number of initiatives throughout 2021, let me give you 2 examples:
- BNP Paribas updated a Data Platform User Charter which commits each one of our data scientists to comply with the highest data security and AI ethical standards
- BNP Paribas deployed an eLearning training module for all employees to create awareness about bias in AI. Furthermore, a specific and mandatory technical training module is being deployed throughout the data scientist community to equip them with best practices to identify and handle bias.
Do you think that feminising the IT professions would achieve the objective of a less biased AI?
Yes, the Group is convinced of this, and in 2020 it launched an IT feminisation programme, as one of its strategic priorities.
Several levers have been identified in this programme, including intensifying external recruitments of women in tech as well as strengthening technical skills development for women. The Group wants to recruit more than 1,000 women in these fields by the end of 2024.
Furthermore, initiatives have been initiated within the Group. For example, Women and Girls In Tech (#WOGITECH) is a collective that brings together women of several generations - whether professionals, students or scholars - around opportunities in digital- and technologies-related jobs.
Finally, the internal and global Women In Cyber network makes it possible to discover the world of women working in the fields of Cyber, Technology and Innovation at BNP Paribas in order to become acquainted with these professions and to encourage job mobility for women.