Broad AI exclusions in insurance policies could create uncertainty, Marsh CEO says
By Manya Saini Oct 15 (Reuters) — A broad-based AI exclusion in insurance policies could create "real confusion and uncertainty," Marsh CEO John Doyle said, calling for a more nuanced approach to assessing risks as the technology becomes increasingly embedded in business operations. Analysts and experts say the rapid development of AI is making it difficult to determine how traditional policy definitions apply to AI-related risks, particularly those involving rogue agents. "We are not trying to poke insurers. We just are really trying to have a thoughtful discussion about how this might change exposures, very limited claim activity to date, and really better understand how these risks aggregate," Doyle said on a post-earnings call on Thursday. AI agents can behave unpredictably and act autonomously, sometimes using access and permissions granted to them, raising questions about whether losses such as business disruption are covered by policies. Exclusions are provisions in a policy that specify which risks, events or losses are not covered. Leading AI developers OpenAI and Anthropic have disclosed that their AI agents behaved unexpectedly, escaping controlled test environments and
By Manya Saini Oct 15 (Reuters) — A broad-based AI exclusion in insurance policies could create "real confusion and uncertainty," Marsh CEO John Doyle said, calling for a more nuanced approach to assessing risks as the technology becomes increasingly embedded in business operations. Analysts and experts say the rapid development of AI is making it difficult to determine how traditional policy definitions apply to AI-related risks, particularly those involving rogue agents. "We are not trying to poke insurers.
We just are really trying to have a thoughtful discussion about how this might change exposures, very limited claim activity to date, and really better understand how these risks aggregate," Doyle said on a post-earnings call on Thursday. AI agents can behave unpredictably and act autonomously, sometimes using access and permissions granted to them, raising questions about whether losses such as business disruption are covered by policies. Exclusions are provisions in a policy that specify which risks, events or losses are not covered.
Leading AI developers OpenAI and Anthropic have disclosed that their AI agents behaved unexpectedly, escaping controlled test environments and carrying out cyberattacks on companies without instruction. Marsh Risk CEO Nick Studer said AI risks currently fall into two categories. The first involves AI as an "elaboration or an evolution of an existing risk or peril." "It will be impossible to unbundle it from the causal pathways to whatever outcome may be. That is really where we were focused on not excluding things which we think would lead to uncertainty, to contention, to gaps in cover," Studer said.
The second category, he said, involves risks that expose gaps or flaws in existing insurance products and services, creating opportunities to adapt them. "Now, we are not certain there is a third category where there is something totally unique to AI that has never been previously contemplated in the insurance market, but we are open to the possibility." (Reporting by Manya Saini in Bengaluru; Editing by Shilpi Majumdar) ((Manya.Saini@thomsonreuters.com; X: manya__saini;)