money.andronezia.com Suka mengabarkan info terkini
Last month Agency Checklist, a US-based insurance news publisher asked ChatGPT, “How will AI change the Property & Accident insurance industry over the next 10 years?” Apparently, ChatGPT told them the following (after listing some of the ways in which artificial intelligence (AI) is expected to change the industry):
“Overall, the use of AI in the property and casualty insurance industry is expected to increase efficiency, improve risk assessment and underwriting, and reduce insurance costs for consumers.”
We decided to ask ChatGPT the same question, but specifically referring to generative AI. The response from ChatGPT was as follows:
“Generative AI, which involves using machine learning algorithms to create new content, is likely to have a significant impact on the Property & Casualty (P&C) insurance industry over the next decade. Here are some potential ways this technology could change the industry:
- Improved risk assessment: Generative AI can help insurers better predict and assess risk by analyzing large volumes of data and simulating potential scenarios. This can help insurers identify potential claims before they occur, which can save them money and improve their overall risk management.
- Enhanced underwriting: Generative AI can assist underwriters by analyzing large amounts of data and creating personalized insurance policies based on a customer’s specific risk profile. This can result in more accurate pricing and coverage decisions, which can improve customer satisfaction and reduce claims costs.
- Simplified claims processing: Generative AI can help insurance companies streamline claims processes by automating certain tasks, such as document review and data entry. This can reduce the time required to process claims and improve the overall customer experience.
- Fraud detection: Generative AI can help insurance companies detect fraud by analyzing patterns in data and identifying potential examples of fraudulent activity. This can reduce the number of fraudulent claims paid out by insurance companies and increase their overall profitability.
- Tailored marketing: Generative AI can help insurance companies create personalized marketing campaigns based on customer data, such as demographics, purchase history and online behavior. This can help insurers better target potential customers and increase their overall marketing ROI.
Overall, generative AI has the potential to transform the P&C insurance industry in a number of ways, from improving risk assessment and underwriting to simplifying claims processing and detecting fraud. As this technology continues to evolve, it will be interesting to see how insurers adapt to these changes and leverage generative AI to stay ahead of the competition.”
This is a bold claim from an AI-powered chatbot that isn’t actually “smart”. Although colloquially referred to as an AI tool, ChatGPT is a GPT (Generative Pretrained Transformer) grand language model that adopts deep learning techniques to generate human-like responses to text input in a conversational way. It is generally accepted that content generated by ChatGPT may not be factually reliable.
In response to this question, ChatGPT is not creatively thinking about new ways in which AI can support the insurance industry. It simply accesses available data and generates answers to the questions asked. Nevertheless, ChatGPT represents the latest technological leap in data analytics and has sparked discussions about the upcoming advances in AI. So when will we get past the AI hype to a position where AI can help insurance companies and insureds? The answer might be “soon, but…”
What ChatGPT can possibly do in insurance
The insurance underwriting process relies heavily on data and statistics. The more an underwriter knows about a risk scenario, the better able they will be to statistically analyze the probability and magnitude of potential loss and price the risk. AI promises to increase the efficiency of collecting, synthesizing, and transmuting large amounts of relevant raw data, into a form usable by underwriters.
Therefore, in terms of risk knowledge and presentation, AI has the potential to be a game changer. We have seen legislative changes reducing the disclosure obligation, for both domestic and consumer line insurers, to the sheer obligation to be careful not to misrepresent. For some classes of insurance, the increased accessibility of publicly available information that AI offers may make the task of disclosure completely redundant.
Much has been said about the use of AI technology in communicating with customers, and the insured is no exception. There is definitely room for AI technology to perform basic interactions with customers when they issue a policy or want to make a claim. This may include creating policy and claims documents, and answering questions about claims applications or processes.
Finally, it is increasingly recognized that the ability to analyze data and detect patterns, as claimed by ChatGPT, can provide insurance companies with useful tools in detecting fraud. This can be a real boon for the industry – not just insurance companies, but also insureds whose premiums could be reduced if insurance companies can reduce the amount they charge based on invalid claims.
All of this may happen very soon. And then there’s the “but”.
ChatGPT limitations (currently)
RGA, a reinsurance company, recently identified the following limitations that ChatGPT and other language models have in the context of insurance. According to RGA:
“These models:
- Trained on a large number of texts from the internet, which may contain bias and may result in the model reproducing biased language and stereotypes
- Lack of common sense knowledge, which can lead to illogical or unrealistic responses
- May struggle with assignments that require knowledge of a particular subject or industry
- Not designed to perform tasks that require reasoning and logical thinking, such as deduction and induction
- Lack of empathy and unable to understand and respond to context emotionally
ChatGPT only accesses the written text when it is trained. It is not capable of leveraging tabular data, charts, or data sources created after its last training.”
With these drawbacks, overhyping the current state of technology risks misleading insurers and insureds.
And why human underwriters aren’t going anywhere anytime soon
Open systems AI technology is unlikely to be able to advise intending insureds what they need and should not disclose to their insurer to fulfill their disclosure obligations, at least in Australia. The reason is that if there is risk-related information that is confidential to the insured, AI will not know about it and will not be able to include it in the consideration of its inquiries. If the matter at issue is “public knowledge”, or is a matter “that is known by the insurance company or in the usual business activities of the insurance company as an insurance company should be known”, including suspected through the use of AI’s own facilities for underwriting purposes as foreseen above, then the information will be accessible to AI but the insured will have no obligation to disclose it (Insurance Contracts Act 1984 (Cth) subsection 21 (2) (b) & (c)). Of course, there is a possibility that AI may progress to the point where it may provide the insured with full legal analysis and advice, informed by all relevant statute and case law, regarding performance of disclosure obligations in the context of the insured’s liability. business, but we’re not there yet.
From an insurer’s perspective, the value of AI technology’s data analytics capabilities is considerable but not decisive. If you ask an underwriter whether their discipline is a science or an art, they will likely say that it is part of both. We hope that insurance brokers and underwriters will not simply abandon human interaction, including with their insured, which can be critical to gaining confidence in those who control or manage the risks insured and to identify so-called “moral hazards” with which insurance worry often.
Can we progress to a world where home and auto insurance policies are filed by the customer, underwritten by the insurer, and generated entirely through an AI technology platform with no human input? Yeah, we hope it might not be far off. Much of the debate in that regard will likely center around insurers and their regulatory oversight of the technology-driven processes and controls needed to make them robust. Life insurance may follow suit relatively soon after, as insurers become increasingly reliant on publicly available health and other data relevant to their potential insureds, to the exclusion of the required disclosures of the humans themselves.
More complex commercial policies, including directors & officers liability insurance and professional indemnity insurance, have traditionally required complex underwriting that is not easily replaced by technology – AI or otherwise. However, much can be done to help. As always, there will be skeptics and detractors, but, like most technological advances throughout history, the overall effect is likely to be positive.
