Insurance professionals worry about AI bias while betting on full automation within two years
Senior insurance professionals across the UK and Europe see AI as essential to their industry's future, yet 87% express concern about bias in AI-driven processes. A survey of 250 professionals by embedded insurance provider EIP found this contradiction at the heart of how the sector views automation.
The same professionals who worry about bias are largely confident in AI's direction. Ninety percent expect their claims administration to be managed entirely by AI within 24 months. All 100% of respondents said AI will have a material impact on insurance.
What else concerns insurers
Beyond bias, professionals cited data security and privacy (23%), regulatory non-compliance (21%), system reliability and errors (21%), and staff resistance (20%) as key worries. Large firms are most anxious about regulatory issues, with 33% naming this as a top concern.
The industry shows least comfort automating claims submissions (40% uncomfortable), followed by underwriting recommendations (39%) and customer interactions (35%).
Transparency as a solution
Thirty-nine percent of respondents said transparent algorithms and decision logs could address AI concerns. This points toward a need for insurance-specific AI tools built with regulatory controls and auditability in mind.
Ross Sinclair, founder and CEO of EIP, said claims decisions require consistency, predictability, and transparency-qualities that probabilistic AI systems struggle to deliver. "The industry needs rules-based decisioning systems configured by insurers, with final judgment left to humans," Sinclair said.
Where insurers are already using AI
Nearly half of respondents (45%) are already using or exploring AI for customer service and chatbots. Risk modelling and data analytics follow at 43%, with claims management (41%), underwriting and pricing (40%), fraud detection (40%), and marketing (38%) close behind.
Cost ranks low as a decision factor-just 10% said it would strongly influence their choice of an AI tool. Integration ease and vendor partnership matter more, each cited by 27% as most important.
Human oversight non-negotiable
Ninety-nine percent of professionals surveyed believe some level of human oversight must accompany AI-driven outcomes. This near-universal view reflects the regulated nature of insurance and the stakes involved in claims decisions.
David Mitchell-Dawson, director of product at EIP, noted that bias exists in insurance regardless of AI use. Rules-based systems avoid this by removing subjective probability assessments from decision-making.
Insurers expect full end-to-end claims automation within 15 months on average, according to the research. For firms not yet moving on AI, that timeline leaves little room for delay.
Learn more about AI for Insurance and how automation is reshaping claims processing and underwriting.
Your membership also unlocks:
