The integration of artificial intelligence into all aspects of the insurance ecosystem is no longer optional, but mandatory for insurers, agents and brokers. There are very real benefits to using AI, but it also comes with unique risks, particularly as threat actors use it in the claims space.
Frank S. Giaoui, PhD, JSD, is an entrepreneur and academic specializing in the intersection of law, economics, and insurance technology. He is founder and CEO of
The following interview has been edited for clarity and space.
1. What are carriers and brokers overlooking as they begin adding AI into their operations?
Implementing AI is a learning process. Artificial intelligence develops with use and its value is incremental. Adding AI is also a low-risk change management opportunity. It provides a useful resource to the team, akin to bringing in a new team member who can communicate effectively with the other members of the team.
2. What are some of the major risks to AI adoption in the insurance industry?
The following are some of the most well-known risks:
· The black box syndrome. When output that comes from an AI system is not explainable and defensible, it may lead to losses in lawsuits where the insurer is challenged to explain its decision not to pay or to pay a reduced amount, for instance.
· Embedded bias in the training data set. If a model is trained using biased information, the model can institutionalize those biases.
· Inadequate controls. An AI system with weak control may be susceptible to breaches in personally identifiable information (PII) and personal health information (PHI).
However, the most important risk will soon be not to use AI, which may be considered a lack of good faith and fair dealing by insurers toward policyholders.
3. How are bad actors using AI against insurers and policyholders?
Bad actors exploit AI against insurers and policyholders in several ways:
· Deepfakes. This involves the use of staged damages,
· False claims. These claims are among the largest contributors to insurance fraud and increasing claim costs. They consist of fabricated, exaggerated or misleading information generated by AI. These claims can range from completely fake or exaggerated to opportunistic or part of a professionally organized scheme.
· Data theft. Malicious actors use advanced methods such as AI-driven phishing, data harvesting, AI hacking and ransomware to specifically target insurance data.
4. What are some of the significant benefits of integrating AI into operations?
Integrating AI into operations
AI can also produce more timely, consistent and accurate decisions in all fields: underwriting, claims and risk management. AI enhances decision-making quality by analyzing large volumes of both structured and unstructured data in real time. In underwriting, AI improves risk selection and pricing accuracy by identifying patterns that humans may overlook. In claims processing, predictive analytics facilitate early triage, severity prediction and fraud detection, which leads to quicker settlements and better reserve accuracy. Additionally, AI enhances consistency by reducing subjective bias and ensuring standardized decision-making across operations. As a result, decisions are made faster, are more reliable and are driven by data.
Insurers can use AI to potentially identify untapped market needs. AI can analyze customer behavior, claims trends, geographic data and market signals to identify unmet needs and emerging risks. This capability allows insurers and businesses to create new products, target underserved segments, and personalize their offerings. Insights gained from AI support innovation, enhance competitive differentiation and enable proactive responses to changing market conditions. As a result, organizations can shift from a reactive approach to a proactive, opportunity-driven strategy.
5. Predictive analytics can play a key role in all types of claims. How should carriers be using it to get accurate information? What kinds of insights can predictive analytics provide? How does this change the claims process?
The following are the major AI use cases for leveraging predictive analytics in the insurance industry.
Early triage and assignment. AI can evaluate incoming claims at the first notice of loss (FNOL) and automatically prioritize them based on such factors as severity, complexity and risk indicators. This process ensures that high-risk or high-value claims are quickly directed to the appropriate adjusters, thereby improving response times and resource allocation.
Fraud detection. AI analyzes patterns and anomalies in claims data to identify potential suspicious activities early. It can detect inconsistencies, fraud rings, inflated losses, and unusual billing practices, thereby helping to reduce claims leakage and financial losses.
More accurate and timely reserving. Predictive analytics enables insurers to estimate a claim's ultimate cost more accurately at an earlier stage. Improved reserving capabilities enhance financial planning, capital management and compliance, while also reducing reserve volatility and unexpected fluctuations.
Automated document summary and data structuring. AI can extract and summarize key information from large volumes of unstructured documents, such as medical records, legal files, adjuster notes and invoices.
Automated report generation. AI can generate claims reports, status updates and analytics summaries automatically using real-time data. This enhances consistency, reduces administrative workload and ensures that stakeholders receive timely and standardized information.
Proactive, fair settlement proposals. AI can assess claim value ranges and litigation risks, allowing insurers to make timely, equitable and data-driven settlement offers. This approach can minimize disputes, manage legal costs and enhance customer satisfaction while promoting fair claims handling.
Pre-litigation and ongoing litigation support. AI aids in analyzing case histories, legal trends, attorney behavior and potential outcomes. It helps prioritize cases that are likely to escalate, supports strategy development, and enhances preparation for negotiations or court proceedings.
Overall reduction in claim cycle time. By automating workflows, improving triage, accelerating analysis, and supporting faster decision-making, AI significantly reduces the end-to-end claims lifecycle. Quicker resolutions lower operational costs, enhance customer experience, and boost overall efficiency .
6.
The alternative name now used is legal abuse or legal system abuse. It is certainly a fairer definition of what is often going on. Some lawyers use the system to enrich themselves. This has no relation whatsoever to the actual damages in a particular case.
The use of these technologies by carriers is still in its early stages. While many discuss AI and explore different options, only a small percentage of industry leaders have fully implemented solutions. Those who have successfully moved beyond the pilot phase are already seeing measurable results. Technology, data, and AI help to prevent legal abuse by enabling the following:
- Early triage and assignment
- Proactive fair settlement proposals or counter anchoring
- Prelitigation and litigation support.
7. Carriers have access to a significant amount of data. Do you find that they are using it effectively to gain actionable insights? If not, what steps do they have to take in order to use their data effectively, and what tools can help facilitate that?
Carriers hold vast quantities of unstructured data, yet it is difficult to process. Many carriers store this data but do not systematically analyze it, so valuable insights related to severity, fraud, litigation and behavioral risk remain hidden. Structured data—organized, tabular and machine readable—is easier for carriers to analyze and operationalize. Carriers that do well in turning unstructured data into structured data gain better actionable insights.
Carriers also have access to a wealth of both strategic (internal) data and industry (external and market) data. However, many have not fully integrated these two data types to generate actionable insights. Although progress has been made, the greatest value comes from combining internal operational intelligence with broader industry signals.
In addition, insurance companies possess large amounts of personally identifiable information, or PII. Although this information is crucial for their operations, it is not always used effectively to generate actionable insights. This limitation is primarily due to privacy concerns, regulatory requirements and governance issues. When managed responsibly, however, PII can provide significant operational and strategic value.
Tools that can help facilitate this data include a secure data lake environment, natural language processing (NLP), machine learning, generative AI (GenAI) and agentic AI.
8. As insurance companies adopt new technologies, implement AI and automate some of their processes, does this make insurance more appealing to the next generation? What types of technology skills do new insurance professionals need to be successful? What opportunities do they have that didn't exist previously?
New technologies definitely make insurance more attractive to the next generation. They allow younger individuals to tap into the accumulated expertise of the retiring generation.
For new insurance professionals to be successful, they do not need to be super-specialized in any specific technology; rather they should train to develop their careers using whatever technologies are relevant to solve their particular use cases. They must be more business savvy than tech savvy. They do have to know what technology does not know. It is more important to have intimate knowledge of the business than the technology itself. Technology is only there to accentuate already existing business acumen.
Technology has enabled the creation of multidisciplinary teams within the insurance sector. New generation professionals from various fields, including engineering, now have opportunities to engage in the insurance industry. For example, software engineers can contribute their coding skills to enhance operations.
9. What excites you the most about the technology opportunities available to the insurance industry today?
Technology is a unique way to add sophistication to a very traditional—sometimes even old-fashioned—industry, potentially attracting more and better talents.
10. What are you watching in the digital technology space? What do you think will be the "next big thing" to change the insurance industry?
The next most important question to consider is for each use case in the insurance industry, what is the optimal balance between human and artificial intelligence? This ranges from having humans involved in high-stakes, low-frequency decision-making (Human in the Loop) to fully automating low-stakes, high-frequency processes (Humans out of the Loop).
