Jane Fraser CEO, Citi
You can barely have a conversation these days without talking about generative AI. At the moment, the technology has limitations and will likely continue to for a while. However, the pace of change is staggering, with the leaps forward happening rapidly.
Citi’s innovation labs have been working on AI — including the kind of generative models that power ChatGPT — for the past three years. In the near term, generative AI will drastically improve productivity. Over the long term, it has the potential to revolutionize all functions across our bank and the industry — changing how we write code, onboard clients, service customers, detect fraud, develop market research and strengthen compliance and controls.
As we integrate generative AI into the firm — and do so in a safe and responsible way — here are the principles that we’re following:
Stay on the front foot: The space is evolving quickly, and we need to be prepared for anything. Some, for example, have surmised that AI personal assistants will be the end of search engines and online retailers as we know it. The impacts on finance will be equally profound. We must be proactive about embracing AI. It’s an essential part of winning in the digital era.
Building a more modern, competitive firm: Finding your unique and attainable AI “why” is critical. At Citi, our work in generative AI is guided by our commitment to transforming and strengthening Citi. We are not chasing some shiny object. The investments and mindshare we’re devoting to generative AI are because we think it will help us accelerate our Transformation efforts and run the bank more efficiently and at greater speeds and scale. This, in turn, will significantly improve how we deliver for our clients, customers and communities.
Amplify the power of people: Generative AI can be a complement to human ingenuity. It can help us become more productive by eliminating routine tasks and freeing us up to focus on higher-order work. Perhaps even more interesting is how it might help us build greater empathy. It may sound counterintuitive, but large language models communicate clearly with their users and never get frustrated. I believe this has wide applicability for many businesses and industries more broadly.
Act responsibly: There are still a lot of questions around regulatory clarity, scalability, guardrails and ethical implications for our clients, and we are in constant dialogue with our regulators about these issues. Justifiably, there has been a great deal of focus on risk mitigation. A rich legislative and regulatory landscape is likely to develop, and we intend to remain at the forefront of compliance with these emerging laws. A focus on safety, though, doesn’t need to be at the expense of action.