At ATB, we are working with leading edge technology with the goal of serving our customers better. Artificial intelligence is one such technology that we’re working with to achieve this goal. AI holds significant promise for the future, but hand-in-hand with the opportunities AI brings are potential uncertainties, and for some, discomfort.
Internally, our ATB team members promise to stand up for what’s right; to put our customers’ best interests first, and to conduct ourselves morally and objectively. That means we seek out emotionally intelligent and ethical people to work with us and why we abide by the ATB Financial Code of Conduct and Ethics.
So, how do these ethical standards translate into the analytical complexities of creating and deploying artificial intelligence to assist us in optimizing our services? Well, we could make Isaac Asimov’s Three Laws of Robotic required reading for all team members, but we’re not about command and control. Rather, our focus is to work alongside team members at ATB to create and instill robust guiding principles as a framework for all of our efforts to use emerging technology in a responsible manner.
This is why we are developing an ATB AI Trust Framework, to ensure our mission is consistent, customer-centric and transparent when working within the computational power of AI.
Our principles for Ethical AI are:
Openness and Transparency Innovation and Social Benefit Ongoing Testing and Safety Openness and Transparency for ethical AI goes for both our team members and our customers. We believe that it is important to be clear about the work we are doing with artificial intelligence so that both team members and customers are able to understand and engage in accompanying dialogue. We want to curate environments that promote diverse, inclusive conversation and allow for all ATB customers and team members to have their voice heard.
Innovation and Social Benefit in developing innovative projects, ATB has the benefit and wellbeing of Albertans in particular, and Canadians in general, as their goal. We’re committed to continuing our investment in enabling technologies and innovative product development; equally important, is the care and attention dedicated to being accountable for the systems we design, and purposeful in our intentions to the communities we serve and belong to.
Ongoing Testing and Safety are critical in everything we do. We bake trust into our development by involving our customers and team members every step of the way, through focus groups, pilot programs and other means of research to ensure the safety of all stakeholders that we serve.
Working on AI technology for the financial industry has certainly helped guide our approach. The regulations of FIs, coupled with ATB’s best practices in: data, security, privacy, and risk all provide the foundations needed to achieve excellence in AI Ethical values. Looking at AI through the lens of the ATB customer keeps us committed to building explainability to our projects, and any implementation of technology has ATB’s values encoded with in it.
In the product development cycle, we are testing structured debate workflows to tease out opportunities to improve on specific ethical dimensions. We believe that our AI Trust Framework is everyone’s at ATB to own and evolve. Transparency, communication, and more importantly, intention, will continue to help us hone in on, and refine, our agile framework as we encounter the changing nuances with this technology.
In the current edition of our AI Trust Framework, we have identified the following ethical dimensions of decision-making related to AI:
- Agency - the ability of customers and staff of ATB to make choices to control their experience working with and within ATB.
- Transparency - the proactive disclosure of the data, methods, and controls that affect the role that AI plays in an ATB customer experience.
- Bias - the effect of (knowingly or unknowingly) making choices that privilege certain groups over others, in ways that are unfair.
- Privacy - the prevention of disclosure or revelation of detail about an individual or group that is invasive or non-consensual.
The framework is still in its early stages, and is built to be agile and iterative as we expand the applications of AI; updating the framework regularly as technology and community evolves. The dimensions considered will also expand and become more robust as we move forward.
We are fortunate to have some of the world’s best AI and machine learning experts right here in Alberta to work with us, as well as help guide our ethical framework through partnerships at the University of Alberta.
Want to learn more about our AI Trust Framework? Stay tuned to ATB Innovation Lab for updates, subscribe below.