Artificial intelligence – How does AI think it can assist insurers?

There has been much discussion recently over the potential application of artificial intelligence (AI) in insurance. To put it to the test we asked AI to draft an article for this month’s edition of The Word. We asked generative AI software to write an article on the impact of AI on insurance.

So, how did it do?

“Artificial Intelligence (AI) is rapidly transforming the insurance industry, from underwriting and claims processing to customer service and fraud detection. As AI technology continues to evolve, it is expected to have a significant impact on the insurance industry in the coming years.

One of the key areas where AI is already making a difference is in underwriting. AI algorithms can analyze vast amounts of data to identify patterns and predict risks more accurately than traditional underwriting methods. This can help insurance companies to price policies more accurately and offer coverage to customers who may have been previously deemed too risky.

AI is also being used to improve claims processing. By automating many of the manual tasks involved in claims processing, such as data entry and document review, AI can help to speed up the claims process and reduce errors. This can result in faster payouts for customers and lower costs for insurance companies.

In addition, AI is being used to improve customer service in the insurance industry. Chatbots and virtual assistants can provide customers with instant support and answers to their questions, 24/7. This can help to improve customer satisfaction and reduce the workload of customer service representatives.

Another area where AI is having an impact is in fraud detection. By analyzing large amounts of data and identifying patterns of fraudulent behavior, AI algorithms can help insurance companies to detect and prevent fraud more effectively. This can result in lower costs for insurance companies and lower premiums for customers.

However, there are also concerns about the impact of AI on the insurance industry. One concern is that AI may lead to job losses in the industry, particularly in areas such as claims processing and customer service. Another concern is that AI may lead to discrimination, as algorithms may be biased against certain groups of people.

To address these concerns, it is important for insurance companies to be transparent about their use of AI and to ensure that their algorithms are fair and unbiased. It is also important for the industry to invest in training and education for employees to help them adapt to the changing landscape of the industry.

In conclusion, AI is already having a significant impact on the insurance industry, and this is only expected to increase in the coming years. While there are concerns about the impact of AI on the industry, there are also many potential benefits, including improved underwriting, faster claims processing, and better customer service. As the industry continues to evolve, it will be important for insurance companies to embrace AI while also addressing the potential risks and challenges that come with it.” 

Whilst the article is rather well written and does cover some key points, there are further areas for discussion that the article omits. 

The article makes reference to customer support and whilst AI is capable of supporting the claims handling process, there can be emotional and psychological components to claims, meaning that many individuals may prefer the support of a human, rather than a chatbot. Equally, chatbots may not be suitable for some customers with specific vulnerabilities. Insurers will therefore need to strike a balance between quick and efficient customer service and ensuring that customers still feel supported and experience a good customer journey.

Before AI can be effectively used in underwriting, the input data will need to match the area of use. It may need to be fine-tuned in order to retain the relevant information and to ensure the best results. Furthermore, whilst the potential for AI to review data may free up time for human employees, AI does not always produce the same outcomes from the same data. It will often provide incorrect outputs, referred to as AI hallucinations, requiring a level of human review. Therefore, at this early stage, the ability to improve efficiency through the use of AI may be limited. 

Without these checks, there is a risk of bias, as referred to in the article, or for customers to be treated unfairly and unequally. As AI uses historic data sets, human review is also necessary to ensure that discrimination is not occurring and transparent models are crucial to provide reasoning behind the underwriting decisions made. However, this can also raise issues, as it may be difficult to assess where discrimination is occurring without collecting data on areas such as protected characteristics.

With AI dealing with large amounts of data, insurers will need to protect this data by taking precautions such as using private instances. Where the software is outsourced, insurers will need to ensure regulatory requirements are met. 

Following the announcement of the UK’s Safety Institute, designed to "understand what each new model is capable of, exploring all the risks from social harms like bias and misinformation through to the most extreme risks", further development is expected in this space to combat issues and allow AI’s potential to be realised. Further development has also been set out in the Bletchley Declaration to vouch for international collaboration on the development of AI and to ensure the identification and understanding of AI risks, whilst creating policies and legal frameworks to manage the area.

The use of AI in the insurance industry is likely to trigger a culture development as staff undergo training on the models and an acceptance of a new way of working. However, used well, AI has the potential to further the insurance industry, potentially allowing for the assessment of risks that are currently uninsurable, whilst AI’s thorough nature can allow it to spot risks that humans fail to.

To hear further discussion of AI considerations for insurers, take a listen to our partner, Tim Johnson’s, interview for the Insuring Cyber podcast.