We sat down with Tim Barnes, Director of Risk and Compliance at Tide and a contributor to The Future of AI Governance and Compliance in Financial Services research report, following the roundtable discussion at its launch in the House of Lords. Tim reflects on how the following points:
1. Generative and agentic AI need a different approach to governance
There is no one-size-fits-all approach to the direction AI is taking. Tide has historically used a data scientist in the second line to validate models, where predictive outputs can be tested against data inputs and past outcomes to judge how likely an event is to happen again. That world looks very different with generative AI and the agentic use cases that now exist. Firms need to work out what governing them actually means, and good looks like.
2. The pace of change is a challenge in its own right
New AI tools are appearing every day. Firms have to judge which is best for a given task, understand the different capabilities of each, and identify the risks that come with them. The technology is moving rapidly, and keeping track of it is a major challenge in itself.
3. AI is harder to oversee than the people whose work it takes on
Many of these roles were historically performed by people, whose behaviour is more predictable. People can be coached and trained, given a procedure and expected to follow the steps. AI works with a power, speed and scale that make it much harder to keep on top of. Where oversight once covered a team of analysts, a single agentic solution can now carry out a large volume of work very quickly.
.png)
