Deciphering AI’s Role in Insurance Raises More Questions Than Answers

At The Professional Liability Underwriting Society’s 2024 cyber symposium in New York City, industry experts delved into the complex relationship between artificial intelligence (AI) and insurance underwriting. Despite AI’s longstanding presence, its rapid evolution, especially in the realm of generative AI, poses unique challenges for underwriters. Garrett Droege highlighted the insurance sector’s sluggish pace in adapting to AI advancements, suggesting that insurers have been aware of AI risks for decades yet have fallen behind in developing comprehensive models to address these risks effectively.

A major hurdle discussed was the ambiguity surrounding AI risks—whether they pertain to cyber or tech Errors & Omissions (E&O) coverages. This grey area complicates the underwriting process, making it difficult for insurers to craft policies that accurately reflect the nuanced risks of AI technology. The existence of AI exclusions in policies was noted, with such exclusions being prepared but not yet widely implemented.

Experts like Jeff Kulikowski pointed out the challenge in distinguishing between technology use within a company and providing a technological service, which directly impacts how risks are assessed and policies are written. This distinction is crucial for correctly identifying and insuring tech E&O versus cyber risks. The conversation underscored the need for insurers to thoroughly understand a client’s operations to evaluate potential losses accurately.

The panelists also touched upon the emerging regulatory landscape surrounding AI, emphasizing the importance of staying abreast of regulatory changes to ensure data protection and network security. Meghan McEvoy highlighted ongoing efforts to work with clients on navigating this evolving risk landscape, stressing the importance of ethical considerations in AI deployment.

Concerns were raised about generative AI’s role in software development and the challenges in attributing liability for issues such as data breaches caused by AI-generated code. This scenario underscores the broader issue of verifying truth in an age where AI can convincingly fake audio and video, posing significant risks for cybercrime.

The discussion ultimately underscored the urgent need for the insurance industry to better define and manage AI risks. While AI presents a daunting array of challenges, from regulatory compliance to liability issues, the experts agreed on the necessity of advancing insurance models to keep pace with technological innovation. The symposium made clear that while the path forward is fraught with uncertainties, proactivity and a deeper understanding of AI’s multifaceted risks will be key to developing effective insurance solutions.