Advancing Responsible & Ethical AI in Financial Services

The growing adoption of Artificial Intelligence and data-driven technologies is poised to transform the global financial ecosystem. They have the potential to improve decision-making, increase efficiency and enhance product and service experiences across the industry.

At the same time, greater integration of AI also has the potential to accentuate certain risks. These include, for example, privacy intrusions, the risk of inaccurate, unfair or opaquely derived outcomes and ineffective accountability structures.

As a financial institution that greatly values innovation and trust, we believe in both the potential of Data and AI to support the financial system’s continued evolution and the need for robust guardrails to manage their responsible use.?

Today, we published our approach to responsible, ethical use of Data and AI, which operates at the top of our wider Data Ethics framework and guides our practical actions:

1. Accountability & Responsibility: Operate effective governance and accountability structures to manage our use of Data and AI, including approval processes, risk management and model review. Take a responsible approach across the AI lifecycle, from initial research and development through to deployment, continuous monitoring and change management. Train employees to understand and apply our approach to the responsible, ethical use of Data and AI, consistent with our culture of accountability. Continuously adapt our approach to respond to the evolving technology, legal and regulatory landscape.
2. Transparency & Explainability: Explain to our clients, employees and all stakeholders, in plain language, how we will use their information. Operate AI systems with appropriate human oversight, making sure that decisions informed by such systems are explainable and justifiable.??
3. Privacy & Security: Employ a “privacy and security by design” approach, including: processing personal information only where we are authorized to do so; upholding the privacy rights of the individuals whose data we use; working to protect data against loss and unauthorized alteration, disclosure and destruction throughout its lifecycle; and avoiding the introduction of new vulnerabilities and security risks from the introduction of untested security solutions that leverage AI capabilities.??
4. Fairness and Accuracy: Identify and mitigate unfair biases throughout Data and AI lifecycles to manage the risk of unfair outcomes. Operate models that are accurate in relation to their intended purpose.?
5. Lawful & Ethical: Process data in a manner consistent with established laws, regulations and our data ethics policy. Design and operate AI and advanced analytics systems to comply with the evolving legal and regulatory landscape, taking into account potential reputational risks. Align our use of Data and AI with BNY’s strategy, principles and code of conduct.

To learn more about how BNY is using Data and AI in a responsible and ethical way, click here: https://www.bnymellon.com/us/en/about-us/technology-innovation/artificial-intelligence/data-ai-ethical-commitment.html