The Intermediary – November 2025 - Flipbook - Page 14
T E C H N O L O GY
In focus
The role oF AI
For high street
lenDers
Dr Paul Dongha,
nsible AI
head of responsible
and AI strategy
at NatWest
“are compliant with material information
requirements and could be vulnerable to
enforcement action.”
Despite this, Timothy Douglas, head of policy
and campaigns at Propertymark, says that the
Government’s drive for material information
packs is “opening up a conversation that looks to
pull together best practice across the industry.”
THE FOUNDATION OF A financial
For Propertymark, a key date is next spring
when the housing department is expected
institution’s operations relies on using IT systems
to come back with proposals from both
that are not only resilient, but also accurate and fair.
consultations.
Artificial intelligence models find patterns in
The industry also expects to see the
data they are trained on, and these patterns form
department’s delayed housing strategy,
the basis for how the model makes decisions and
which in July it pledged to publish “in the
recommendations. However, in exceptional cases,
coming months.”
there is a risk of algorithmic bias appearing in
AI models.
As a bank, we have to ensure that we don’t allow
Douglas says: “These documents should set
out a roadmap for reform. They will give the
industry a clearer direction of travel.”
this to happen. The financial decisions we make
– relating to everything from mortgage approvals
Starting the sprint
to loan decisions – impact people’s lives and
Many across the industry have long heard cries
their futures. We therefore aim to take the necessary
from the Government and regulators about the
steps to ensure our AI models make correct and
need to drive radical change in the housing
fair decisions.
market – only to see tinkering at the edges.
Central to our approach is our AI and Data Ethics
Code of Conduct. The code outlines our fundamental
At the very time of writing, however, the FCA
has commenced a three-month ‘TechSprint’ with
principles regarding the ethical use of AI and
home loans industry firms to “simulate and test
reflects our aspirations to foster responsible and
data sharing,” suggesting that the time may well
transparent practices.
have come for real movement to start.
It also contains principles that govern how we use
and process customer data and how our AI systems
The regulator says: “The aim of the Mortgages
TechSprint is to explore and test how open
are developed and used. The code gives all our AI
finance and smart data can enhance the
engineers a framework through a set of high-level
mortgage journey by developing practical
principles that they need to work within.
It works towards ensuring: that our AI systems
solutions that improve access, affordability, and
consumer understanding, while informing
are subject to human oversight, and that they
the FCA’s regulatory approach and supporting
respect and promote human agency; that they are
broader Smart Data Accelerator initiatives.”
technically robust, resilient, and safe, so we prevent
The test will be run by tech firm Raidiam,
unintentional harm to individuals; that we comply
which has pioneered moves toward open finance
with privacy and data protection laws; and that
and open insurance in Brazil.
decisions or predictions that they produce can be
The message is clear: the mortgage and
explained to customers and are free from unfair bias
homebuying process can no longer rely on
or discrimination.
incremental tweaks or isolated pilots.
Cases deemed high risk, referred to as ‘edge
With Government consultations underway,
cases’, are formally discussed in our AI and Data
new legislation enabling smarter data use, and
Ethics Panel, which evaluates the ethical concerns
the FCA actively testing open finance solutions,
based on the principles in the code and makes
the conditions for genuine transformation are
recommendations as to how potential ethical risks
finally aligning.
should be mitigated.
Each member on the panel represents our
different stakeholder groups – including customers,
Yet technology alone cannot solve the problem.
The prerequisite is trusted, accurate, shareable
data – and a framework that gives lenders,
colleagues, communities, shareholders and investors,
brokers, conveyancers and consumers confidence
as well as future generations.
in how that data is used.
Without this foundation, automation could
introduce as much risk as it removes.
12
The Intermediary | November 2025
●
