Nearly two-thirds of European insurers are already using generative AI (Gen AI), according to a report published by EIOPA.
Most undertakings are nevertheless still at a proof-of-concept stage, indicating a considerate, controlled roll-out of the technology as well as significant potential for future growth.
Insurers said their main motivation for implementing Gen AI tools was to improve efficiency and cut costs, but more than half of the respondents also cited efforts to enhance the customer experience and the desire to improve decision-making processes as key drivers.
Although a large number of insurers are eager to tap the expected productivity gains of Gen AI tools, they are taking a cautious approach by focusing on internal processes first and maintaining strong human oversight. The majority of the reported use cases (64%) target back-end productivity tools such as data extraction from invoices, audio recordings or medical reports, content generation for emails, contracts or marketing materials, or coding and underwriting assistants.
In addition, 36% reported the development of customer-facing Gen AI applications such as voice- or chatbots. While most of such applications are at proof-of-concept stage, autonomous Agentic AI applications are expected to have a greater impact in this area over the medium term.
Respondents also provided input on the key challenges that hinder or slow down their adoption of AI systems. They identified data privacy and security concerns, potential issues with regulatory compliance and the lack of sufficiently skilled staff as the main hurdles to successfully implementing and managing Gen AI systems.
Hallucinations, or inaccurate outputs, was the top-cited risk connected to Gen AI systems, followed by cybersecurity risks and data protection issues. Responses also highlighted that insurers rely heavily on third-party providers, opting to purchase off-the-shelf solutions or build on top of pre-trained models. Respondents also pointed using sectoral rules, the Digital Operational Resilience Act (DORA) and the AI Act to address the reliance on third-party providers.
These risks and their potential implications underscore the need for insurers to adapt existing governance and risk management frameworks to address AI-specific challenges. To this effect, 49% of undertakings in the sample have already developed dedicated AI policies, up from only a quarter in 2023.
Petra Hielkema, chair of EIOPA said: “The findings of our survey show that European insurers have been carefully considering how Gen AI could benefit their businesses and consumers. While several solutions are already being implemented, even more is in the pipeline. The fact that this scale-up is happening gradually, with strong human oversight, and alongside updates to risk management frameworks, is important in order to adopt AI responsibly. That is what we want to see going forward. We will be monitoring developments on the market with national supervisors.”
The report was based on responses from 347 undertakings across 25 countries.
