THOUGHT LEADERSHIP
claims for car accidents, the chatbot can provide them with information about car insurance policies and accident prevention tips.
answer is to deploy LLMs configured with restrictive context output. This will make the behaviour of your language model more predictable and avoid liability issues. Ethics The use of conversational and generative AI in the insurance industry also raises important ethical concerns, including the potential for bias, and the need for transparency and accountability. “How do we ensure that AI implementations are ethical?” “How do we take action to mitigate ethical risks?” and other burning questions were answered in the latest publication of Harvard Business Review on ethical considerations of using AI. But the golden nugget here is that to ensure that these technologies are used for the greater good within insurance organisations, companies must partner with providers that promote ethical considerations and responsible development and have an AI code of ethics in place. Data Privacy Generative AI tools can pose risks to data privacy in several ways such as data breaches and unauthorised data sharing. Insurance companies possess personal customer information and must ensure its security. Therefore, they are hesitant to implement tools that cannot safeguard digital privacy and throw data out into third-party APIs. The solution to this is to deploy Generative AI on-premise with a provider that ensures compliance with PII, GDPR and internal policies, for example, through validation with IBM Cloud for Financial Services. Are Bots Going to Completely Replace Humans in the Workplace? Although there is concern that bots may replace humans, AI is simply a tool to help humans improve their work. For instance, AI-powered bots can help insurance debt collectors deal with the emotional challenges of collection calls. In addition, consumers who are embarrassed to discuss their debts with others prefer to speak with a robot rather than a person. This way, AI can reduce emotional strain for the agent and the customer, streamline experiences and improve call centre efficiency. However, human decision-making and compliance with regulations will always be necessary.
CONTINUED
Empowering Employees with Knowledge
Apart from customer engagement solutions for insurance, 2023 brings other advances. This year Generative AI innovations revealed the potential for a radical transformation of business data querying. Enterprises can now use large language models to facilitate what we call “chatting with data.” What this means is that businesses can deploy custom ChatGPT- like search bots tailored to their specific internal needs to enable document search. The growing interest in document search bots among insurers and other big players can be explained mainly because of its apparent benefits. These bots understand company- specific data and terminologies and provide accurate and relevant responses. Employees can ask questions and get information in a conversational format from different data sources spread across the business. That could be anything from a PowerPoint to a Word document to an Excel file. Coveo’s study found that the average employee spends 3.6 hours daily searching for information – one hour more than last year’s report. The frustration caused by not being able to find information easily contributes to burnout, but thanks to the new solution of generative AI for business this risk can now be significantly reduced. What Are Emerging Concerns With AI in Insurance in 2023? Digital innovations of AI in the insurance industry allow organisations to be more effective and efficient in the ways they interact with customers and corporate data. Yet, as with any innovation, the opportunities presented by evolving AI come hand-in-hand with new and growing concerns. Output Unpredictability How do companies use AI large language models in such a way that they control the output and don’t get what’s called a “black box effect” discussed in one of the Forbes publications? Particularly if these models are used in a customer-facing setting, where compliance is paramount, and we want to ensure there are no reputational risks. The
How do we ensure that AI implementations are ethical?
?
How do we take action to mitigate ethical risks?
48
Powered by FlippingBook