AI in Banking Shifts From Performance to Accountability
AI in banking is moving from experimentation to accountability. What stood out in this conversation with Metro Bank is a clear shift in mindset: performance alone is no longer enough. Ben Guy, Head of Reporting and Impairment at Metro Bank, highlights a more mature approach, one where model performance, operational efficiency, and governance are treated as a single system, not separate concerns. That matters. Because the next wave of AI in risk and finance won’t be defined by who builds the most sophisticated models. It will be defined by who can trust them, scale them, and embed them into real decisions. A few themes that come through strongly: 💠 Reliable modelling is foundational, but only valuable when it holds up under scrutiny 💠 Governance is not a constraint; it’s what enables confident adoption 💠 AI works best when it augments expert judgement, not replaces it This is the direction of travel across financial services: from models in isolation → to decision systems with accountability built in If you're working in risk, finance, or model governance, this is worth a watch. Curious to hear how others are approaching this balance between performance, trust, and real-world decisioning.
AI Finds Patterns; Humans Decide What Matters
AI can surface patterns, but it cannot decide which ones matter. That responsibility remains human.
New PhD Tackles GenAI Governance in Finance
Generative AI is moving fast into financial services, but model management, explainability, and governance are still catching up. The teams at Nationwide and Cardiff shared this funded PhD with me, and I wanted to share it with my network. The...
AI Arms Race: Fraud Advances Faster Than Defenses
Financial crime isn’t slowing down. It’s accelerating and AI is changing both sides of the battle. Fraudsters are scaling faster, automating attacks, and exploiting gaps between systems, teams and governance. The question is no longer whether fraud is evolving. It’s...
AI Agents Become Profit Drivers, Demand New Accountability
Agentic AI is starting to change something fundamental in financial services: who actually drives profit and loss. We are moving beyond models that inform decisions. AI agents are beginning to price risk, approve loans, detect fraud, handle claims, and manage...
AI Agents Need Product‑style Lifecycle Management, Not One‑offs
AI agents rarely fail where most teams expect them to. They don’t fail in development, where everything is controlled and tested. They fail later, once exposed to real-world variability. I see this repeatedly. An organisation builds an agent, the demos...
Future‑proof AI: Learning Ability Outweighs Launch Accuracy
Why the most valuable AI systems are not the most accurate ones today, but the ones designed to learn tomorrow In the early days of enterprise AI, success was measured in a single moment: the model launch. A team would...
AI Flags Alzheimer’s Risk Two Years Before Diagnosis
Earlier diagnosis can change the trajectory of Alzheimer’s. Yet thousands of people in the UK still live with dementia for years before receiving a formal diagnosis. The signals are often already present in clinical records, they are simply fragmented across...
AI Should Amplify, Not Replace, Human Thinking
True progress in AI is not creating systems that think for us, but creating systems that help us think more deeply
AI Teaches Us About Bias, Complexity, and Knowledge Fragility
As AI learns from us, we must also learn from it, about bias, complexity, and the fragile nature of knowledge
Embedding AI Requires Culture, Process, Not Just Pilots
From isolated experimentation to embedded capability, culture, process, and incentives as the true transformation levers In boardrooms across industries, the same slide appears with reassuring regularity: a grid of successful AI pilots. A fraud model that lifted detection rates. A...
Data’s Objectivity Is an Illusion; Human Choices Shape It
Data is objective only in appearance. Behind every dataset lies a human decision about what to measure
Algorithms Reflect Our Worldview—Shape Them Responsibly
Every algorithm we build carries a piece of our worldview. Responsibility means shaping that view with care
Human Creativity Thrives Beyond AI's Logical Limits
AI will never replace human creativity because creativity begins where logic ends
Never Stop Questioning Models, Or They'll Mislead You
The moment we stop questioning our models is the moment they start misleading us
