How Citi Is Balancing AI Ambition With Safety

Jon Lofthouse from Citi talks candidly about just how hard it is to predict where AI, and especially large language models, will take the financial services industry. He admits that even a year ago he wouldn’t have guessed we’d be where we are today. For him, that uncertainty isn’t a failure of planning, it’s a sign of how fast both the world and the underlying technology are moving.

Lofthouse emphasises that as Citi leans into these tools, risk management remains non-negotiable. He stresses the importance of staying “safe and sound” and keeping strong guardrails and checks in place as the bank experiments and scales up its use of AI. The message is that progress won’t be reckless or purely hype-driven; it will be controlled, supervised and aligned with regulatory and safety expectations that matter in financial services.

At the same time, he’s clearly excited about the upside. Lofthouse talks about the potential to modernise core systems, supercharge what software engineers can deliver and rethink how the bank serves clients. He points to opportunities to automate more of the operational processes that sit behind the scenes in a bank, not just to cut manual work but to improve quality and consistency.

Citi closes by reflecting on how quickly large language models have improved, and what that implies for the near future. If it’s already this hard to predict a one-year horizon, they suggest, then the next few years could unlock possibilities that are difficult to even imagine today. The overall tone from Citi is a blend of ambition and caution: move fast enough to capture the benefits, but not so fast that you compromise on safety, soundness or trust.

The post How Citi Is Balancing AI Ambition With Safety appeared first on FF News | Fintech Finance.