IRSG Advice on EIOPA’s Opinion on Artificial Intelligence Governance and Risk Management

IRSG INSURANCE AND REINSURANCE STAKEHOLDER GROUP Advise on the EIOPA’s Opinion on Artificial Intelligence Governance and Risk Management IRSG-25/08 12 May 2025 INTRODUCTION We are grateful for EIOPA's invitation to comment on its Opinion. The interest in the subject has provoked intense debate within the IRSG, to the extent that the space allotted to respond to each of the questions raised has been limited. For this reason, we have uploaded a shortened version via the link on the official EIOPA website, but we are submitting this longer version so that the reasoning behind each response can be better understood. 1. QUESTIONS TO STAKEHOLDERS Q1 - Do you have any comments on the context and objectives of the Opinion? We appreciate EIOPA’s initiative and its stated objectives. Although most of the AI systems deployed by insurance undertakings are neither prohibited nor classified as “high-risk” by the AI Act, advanced AI systems extensively used in core insurance functions such as underwriting, pricing, and claims management may generate or increase risks that require proper oversight, taking into account to what extent the business of an insurance undertaking relies on AI systems and focusing in risk prevention (e. g. “ethical AI by design”). Also, we agree on the choice of issuing a high-level opinion addressed to supervisors, instead of detailed guidance addressed to insurance undertakings, and on the need to not add...