セレント、シカゴで生成AIシンポジウムを開催

Celent held a full day event for insurers focused on Generative AI (GenAI) and Large Language Models (LLMs) in the insurance industry last Wednesday in Chicago. It was a combination of panels, outside speakers, and Celent presentations. It really was a phenomenal event.

After an introduction by Karlyn Carnahan, Andrew Schwartz presented Beyond Human Intelligence: Unleashing the Power of Large Language Models in Insurance. He began the session level setting how GenAI and LLMs work and how they differ from artificial intelligence (AI). He then brought a perspective into GenAI and LLMs could be used in insurance. While a recent Celent poll showed that most insurers thought customer service was the most logical place for GenAI and LLMs to be implemented, Andrew presented five opportunities for LLMs in P&C insurance: personalizedmarketing​, enhancedclaims​ processing​, frauddetectionandprevention​, productinnovation​, and risk assessment​. He also presented instances where insurers had implemented LLMs in their business and some vendors with GenAI/LLM solutions across the insurance value chain.

Andrew also presented the risks associated with LLMs, highlighting adverse outcomes like biases and unethical behavior, hallucinations, and false outputs, and falling victim to nefarious actors. External threats like regulatory or intellectual property violations were also discussed. Andrew lightly discussed strategies for tackling the risks knowing that a panel would discuss it more fully after his session.Andrew’s ended his session with some recommendations for how an insurer could leverage Generative AI but still manage risks.

Donald Light moderated Regulating the AI Revolution which discussed the current and potential future legal, regulatory and ethical landscape surrounding the use of Generative AI in the insurance industry. The panelists included:

• Rick Borden - Cybersecurity and Privacy Partner, Frankfurt Kurnit
• Maria Kokiasmenos, VP, Associate General Counsel, The Hartford
• Lionel Weaver, Counsel, Faegre

The discussion touched upon how the regulation of GenAI, which is still being developed, differs from established types of AI that are in use today and whether GenAI would fall under the AI umbrella or would require a different regulatory approach or just a different approach.The consensus was that GenAI would likely still fall under the broader AI regulatory umbrella, albeit with specific guidelines tailored to its nuances. Donald also asked if the regulation for AI/Gen AI would move to a to a “risk management oriented regulation” and if so, how does that approach would differ from long established methods of regulation. The panelists agreed that a risk management approach was likely. The recent regulatory initiatives in Colorado and the July NAIC Bulletin brought forth an interesting discussion about the approach, how other states might react, and how much insurers should look to these documents as a basis for creating regulatory response. The panel continued with discussions around the major governance and risk management structures and processes insurers should have in place, the convergence of regulation of AI/Gen AI, privacy, and cyber, and lastly how insurers might modify their governance and risk management structures and processes. After a conversation around regulatory issues related to third party providers of data and/or Gen AI and the factors insurers should consider when working with third parties, Donald ended the session with a question around the ethical issues of using Gen AI in particular and of all AI in general (that is issues not covered by statutes or regulations).

The highly anticipated pre-lunch session, Innovating Insurance: Generative AI Panel Discussion featuring AWS, Google, and Microsoft did not disappoint! Moderated by Keith Raymond, the panel featured:

• Frank Neugebauer, Insurance Industry Solution Manager, Google
• Aarti Gupta, Insurance Industry- Head of Business Development for NAMER, AWS
• Jim DeMarco - Industry Digital Strategist, Insurance Strategy Lead – Microsoft

Keith probed the panelists about the impact, challenges, and opportunities that GenAI presents in revolutionizing insurance processes, customer engagement, risk assessment, and fraud detection. Each vendor described how their organization is currently thinking about the opportunity for GenAI and its application for insurance. They also provided examples of how GenAI could improve processing efficiency for insurers or their own companies.The discussion continued with the skills needed for the future to support the development and maintenance of LLMs. Continuing on some of the previous panel’s conversation Keith asked about the technical, security, regulatory and risk challenges insurers need to consider when implementing an LLM solution, and what insurers be thinking about today in their strategic planning to be prepared for the application of LLMs.

A particularly noteworthy moment arose when a panelist emphasized the potential benefits of big tech players like AWS, Google, and Microsoft collaborating on GenAI. The idea of pooling insights and best practices to collectively enhance the positive impact of GenAI in the insurance sector resonated with the panel, as evidenced by nods of agreement from the other participants.

During lunch Cognizant presented the results from a Cognizant survey, Understanding AI: The Opportunity for Value Creation. With only 20% of banks and insurers already implementing AI in their company, Cognizant sees the real potential for AI in 2025 or later with the greatest impediment being conflicting projects and the state of data in the companies. Like Celent, Cognizant believes the greatest short term impact will be in customer experience.

Nate Golia opened the afternoon with Unveiling Risks, Overcoming Challenges, and Mastering Governance in Insurance, a discussion about ethics. Nate’s panel included

• Jaymin Kim; Senior Vice President, Emerging Technologies, Cyber Practice US & Canada, Marsh
• Franklin Manchester, Global Insurance Strategic Advisor, SAS
• Gina Cristallo - Vice President & Chief Ethics Officer, New York Life Insurance

Picking up from Donald’s panel on regulation, Nate probed whether discussions about ethics should go beyond whether a practice is illegal and if insurers should be thinking about risk management from a reputational standpoint as consumer reaction is likely to come faster than, and possibly influence regulation. The panel discussed how to incorporate AI literacy into their training and what business and tech leaders should be communicating right now about Gen AI in terms of how and when to use it and how to protect company data.

The symposium ended with Showcasing the Power of Generative AI in Action by Market-leading Vendors​. The showcase featured five vendor solutions that offer varying levels of GenAI along the insurance value chain. The five vendors included: BetterView, DAIS, Digital Owl, Majesco, and Appian.

At the end of the day, it was clear that the use of generative AI in insurance is still in its early stages. While there are some concerns about the ethical implications of using AI in insurance, such as the potential for bias and discrimination, many believe that the benefits of this technology outweigh the risks. As the technology continues to evolve and improve, it is likely that we will see more widespread adoption of generative AI in the insurance industry – leading to a greater focus on governance. GenAI capabilities range from predictive modeling, code generation, and risk assessment to delivering tailored customer experiences. As the pace of advancements accelerates, insurers will need to continue to stay informed and educated on the topic.

If you are interested in talking to Celent about their GenAI and LLM research, reach out to Chuck Smith at CSmith@celent.com! We look forward to talking to you!