Insurance sector must avoid 'built in bias' with AI, MPs hear

Responsible and regulated use of AI can help the insurance industry alleviate the impact of climate change and stem the widening protection gap, a leading academic has told parliamentarians.

Dr Simone Krummaker, associate professor of insurance at Bayes Business School, said the transformative nature of the technology’s impact on the sector calls for an industry-wide response.

Addressing the All Party Parliamentary Group on Artificial Intelligence, she said: “The widening protection gap – including here in the UK – is the test that matters. If AI cannot help more people and businesses secure affordable, appropriate cover, then the sector has missed the point. Used well, AI gives us sharper sight of risk, helps prevent losses before they happen, and directs capital towards resilience. But it only works if it is governed well, with fairness designed in and demonstrated in real outcomes.”

While the technology will help the sector meet rising consumer expectations, she continued, AI tools and systems must be developed and implemented responsibly. She cautioned that the industry must build and retain public trust as it exploits the benefits of AI. Professor Krummaker said:

“As digitalisation accelerates, automation must not cement bias. Customers should understand how and why decisions are made. That requires explainable models, clear communication and outcomes that can be challenged and reviewed. When people understand why something happened, they are far more likely to trust it.

“The same principle of fairness applies to underwriting and pricing – areas that are already deeply influenced by data and algorithms. My colleagues at Bayes have explored how these systems can be made fairer and more transparent. We’ve found that algorithms, if left unchecked, can actually amplify inequalities rather than reduce them.”

“This is not simply a regulatory concern; it’s a matter of legitimacy and public trust. The sector’s credibility depends on customers believing they are treated equitably – not as data points, but as people.”

“AI raises expectations for service and speed, but it also raises the bar for responsibility. The best deployments are not the flashiest; they are the ones with clear decision boundaries, human oversight where it counts, strong model governance and honest explanations around claim decisions and premia that people can act on. In other words, ethics is not a slogan we recite at the end – it is the operating system that keeps trust alive.”

She emphasised the need for companies to work with each other and with policymakers and regulators to ensure the benefits of AI are rolled out responsibly.

“None of this can be achieved by a single organisation. Closing the protection gap and modernising insurance is a team effort: insurers and reinsurers, managing general agents and brokers, technology providers and researchers need to work alongside regulators, public data holders and local authorities. When we share data responsibly, align on data protection assurance, and learn openly from what works and what does not, good practice scales quickly and benefits everyone, not just the largest firms.”

A partnership model between the public and private sectors, she said, should include shared catastrophe and resilience datasets, proportionate AI assurance expectations, and incentives for innovative risk-sharing that keep cover available and affordable in at-risk communities.

That approach would support a positive workplace culture that gives people in claims teams and underwriting rooms the skills, confidence and curiosity to ensure that AI benefits those the sector serves, Dr Krummaker said.