The GenAI impact on insurance: Part 2
Aug 2024
Michael BewleyAI tools and aerial imagery can enable more precise underwriting and better risk management.
Aug 2024
Michael Bewley)
In Part 1 of this blog, I covered how AI tools like GenAI currently benefit the insurance industry by helping them understand property risk better and tailor their policies to reflect the actual conditions and potential risks associated with each property. But how does the combination of AI tools and aerial imagery play into more precise underwriting and better risk management? How can the use of GenAI remain fair, transparent, and equitable for all parties involved?
AI & aerial imagery
At Nearmap, one of our core uses of AI involves analysing aerial imagery to assess property risk — also known as computer vision. We’re building an accurate, continually updated model of the physical world to identify risks, trends, and changes.
For instance, Nearmap AI can already process and interpret millions of square kilometers of aerial data, identifying details that are crucial for accurate risk assessment. As another example, Betterview, our comprehensive insurance platform, can detect conditions of buildings, risk factors like solar panels or swimming pools, and environmental hazards, and then either flag or automate risk profiles.
The result is insurers get a true view of risk and reduce the guesswork and generalisations of traditional insurance assessments.
GenAI also plays a role in enriching our technology. For example, a large language model chatbot is leveraged in the Betterview platform, where an insurer can query the risk profile of a property and receive an answer in seconds. While it’s an exciting trend, the focus on rock-solid AI Systems means outcomes are more important than the precise technique used.
So, GenAI has a place, but there’s much more to making real-world AI. Nearmap AI is a holistic system that goes from pixels to risk models, and a complex, carefully designed set of algorithms and systems between. It can assist with:
- Detailed risk detection: AI systems can be trained to detect various conditions and features of properties that are critical for assessing their risk levels. For example, AI can automatically identify building characteristics like roof integrity, age, and materials used — which can impact the vulnerability of the property.
- Environmental hazard analysis: AI algorithms extend to detecting environmental hazards that could impact properties. This includes identifying overgrown vegetation that might pose fire risks or unstable land that could lead to subsidence. These insights are crucial in regions prone to specific types of natural disasters, allowing insurers to evaluate potential risks more comprehensively.
- Impact on insurance policies: Having a thorough understanding of property risks helps insurers to customise policies that better reflect the actual conditions and potential risks associated with each property. This ability to tailor policies based on detailed, property-specific risk assessments significantly reduces the reliance on broad actuarial tables and guesswork.
- Improve underwriting precision: The use of AI systems in analysing aerial imagery leads to more precise underwriting. Insurers can adjust their policies and pricing to more accurately match the risk level of each property, improving their risk management strategies. This level of precision not only benefits insurers by reducing unexpected payouts but also benefits policyholders by not overcharging them for lower-risk properties.
As GenAI, and AI Systems more broadly, continue to transform the insurance industry, what does the future hold for this technology? With its potential for continuous improvement and innovation, we can expect even more advancements in risk assessment, policy customisation, and decision-making processes.
The future of GenAI for insurance
As insurance continues to adopt AI technology, the industry must ensure that models are transparent, equitable, and free from the biases which influence decision-making. This is done by maintaining rigorous standards on the data used to train these models. The security and privacy of data, adherence to regulatory requirements, and ethical considerations in AI deployment are paramount. Opportunities afforded by GenAI in interaction with text, enhancing images, and learning patterns from large quantities of data are no exception – the same rules apply.
It is the insurers’ responsibility that artificial intelligence is used responsibly, improving customer trust in AI-driven processes. This includes:
- Transparency and fairness: Transparency is a central tenet in the responsible use of GenAI. AI systems must be designed in a way that their decision-making processes are understandable to users and stakeholders. Critically, when the AI System is uncertain, it must communicate that to a customer. This is specifically a challenge for GenAI-based systems that must be addressed. It also needs to be transparent about the provenance of the data. Where did it come from? How was it captured? Precisely which data sources did the algorithm rely on? This challenge increases in importance as the number of data sources increases, and GenAI algorithms are able to reason about text inputs more readily. This openness helps in building trust and also makes it possible to monitor AI decisions and improve the model’s fairness and accuracy.
- Equitable AI decisioning: AI models should be free from biases that can influence decision-making. This is coupled with transparency, as the context of how and why the decision was made, and the evidence on which the output was based must be considered when analysing bias. It’s vital to examine and understand the way the datasets used for training these models were created, to avoid inequalities or new biases.
- Data integrity and privacy: Maintaining rigorous standards on the data used to train AI models is fundamental. This involves verifying the data’s accuracy, relevance, and currency. AI data must also be sourced ethically, respecting individual privacy rights and adhering to data protection regulations. This can be challenging without control over the end-to-end process by which data was captured, processed, used in training, and deployed to provide an answer on a particular property.
- Regulatory compliance and ethical deployment: Insurance is a heavily regulated industry, and AI Systems, including applications of GenAI, must comply with all relevant laws and regulations. This includes regulations around data use, privacy laws, and rules governing AI decisions. Navigating this regulatory landscape requires a proactive approach to compliance and an ongoing dialogue with regulatory bodies to align AI applications with legal standards.
AI systems were already in the process of revolutionising the insurance decision-making process. GenAI adds some new elements, risks, and opportunities to consider. As AI technology continues to advance, we can expect even more benefits, such as faster claims processing and fraud detection. Insurers who embrace this technology will have a competitive edge in the market, providing better services to their customers and making smarter decisions that benefit both sides.
Want to learn more about GenAI for insurance? Contact us today to try it for yourself.