Insurance In The Age of AI: Smarter, Faster, Or Kinder Yet?

Artificial Intelligence (AI) is not new in the insurance industry—the industry that needs to think ahead, create many scenarios to predict potential and reduce risks. So, clearly insurance companies cannot stay away from the AI dominance.

What is AI in insurance?

AI in insurance combines AI, automation, and other innovations to enhance coverage and smooth out service delivery.

Simona Scattaglia, Global Insurance Technology Lead and Partner at KPMG in Italy, highlights the promise and wide impact: “CEOs recognize that AI and generative AI are technologies with huge potential for their business, because they touch on so many core aspects of what insurers do.” (1)

AI in insurance use cases

It’s clear to see how AI causes impacts on the creative field, where those machine learnings consume big data and generate thousands of ideas in the blink of an eye. So, specifically for the insurance industry, what will it transform?

Here are key takeaways from the IBM article (2):

– Claims processing: AI can automate many steps in the procedure, such as initial assessment, document capture, or even analysis before conducting payment authorization for simple claims. This speeds up the process (it would take up to 15 days) and reduces the administrative burden for insurers with fewer paper documents and, finally, improves customer satisfaction when they get the coverage much faster.

– Fraud detection: Insurance fraud is a costly problem. AI-powered systems can identify suspicious and unusual patterns in claims data that might indicate fraud, helping insurers cut losses.

– Risk management: AI can analyze data from various sources to identify potential risks. For example, analyzing the customers’ health records could help insurers proactively be aware of the risk of severe sickness or genetic condition. Accessing data from cars or homes through smart devices can also provide insights into risky behaviors. These statistics help insurers’ agents develop more tailored insurance products and more accurate, personalized pricing.

– Underwriting: AI algorithms can consume and analyze vast amounts of data, including traditional insurance records, credit scores, daily activity, and even data from connected devices, to assess risk more accurately and efficiently than traditional methods.

– Customer experience: AI-powered chatbots or virtual assistants can answer customers’ questions immediately, especially for the frequently asked ones. Then, it could guide them through policy step by step, which aims to improve overall satisfaction

Challenges for AI in insurance

Data privacy and security

AI systems require vast amounts of personal and non-personal data to work out. This raises significant concerns about data privacy and the security of sensitive information (medical history, health records, and home address) of the customers who signed the contracts, whether for themselves or others.

Meanwhile, the insurance industry is on high alert from cyber-attacks, according to Google’s Chief Threat Analyst (3). Reported by the BBC, in the middle of July this year, hackers had stolen 1.4 million customers personal information from insurance firm Allianz Life. After some attempts, the company was still unable to specify how many people had been affected (4).

Transparency and explainability

Most AI models are often known as “black boxes” because not many people understand how those models work and provide answers. This lack of transparency can be a major issue, especially when a consumer wants to know why their claim was denied or why their package is higher than someone else’s if there’s no support from human agent.

Discrimination and biases

It is assumed that AI neutral mind will eliminate biases and discrimination, leading to smarter decisions. But it’s not true.

Recently, they found that AI does create inequality, developing false assumptions according to gender, race, age, profession, etc. AI models are only as good as the data they are trained on; data that is used to train AI needs to be clarified, centric, structured, and non-biased.

So, specifically for the insurance sector, once AI proceeds with biases and discriminations, it can easily decide to refuse a claim or overprice a case after its algorithm works out. Then, it all affects the decisions of insurance companies, leading to unfairness for certain groups—who need medical aid.

According to the Guardian (5), insurance coverage denials have risen in recent years in the US, driven in part by automated algorithms powered by AI—and some recently launched artificial intelligence tools may fight back by generating automatic appeals. One of the lawsuits alleges that Cigna denied more than 300,000 claims in a two-month period, which amounts to about 1.2 seconds for each physician-reviewed claim.

Regulatory and compliance issues

Not an overstatement that insurance is the most heavily regulated industry (6). Firstly, it’s strictly followed by international regulation such as the General Data Protection Regulation (GDPR) or the Health Insurance Portability and Accountability Act (HIPAA) for data protection and the Payment Card Industry Data Security Standard (PCI-DSS) to ensure financial security.

The next thing that cannot be ignored is the local requirements for each industry in order to streamline the whole operation. For example, company needs to abide by specific frameworks, in the US by the National Association of Insurance Commissioners (NAIC) or the Insurance Regulatory and Development Authority of India (IRDAI).

Different perspectives on AI in insurance:

AI is rapidly reshaping the way people work. Now, in the insurance landscape, it offers powerful benefits while raising concerns that cannot be ignored.

– For insurers: AI offers the potential for increased efficiency, eliminated manual tasks, improved risk assessment, and enhanced fraud detection. However, insurers also need to consider the investment costs associated with implementing AI technologies, the need for data security and privacy, and the potential ethical implications of using AI in decision-making.

– For consumers: AI’s “black box” process also brings up questions about trust and biased decisions. Their key concerns include data usage and consent, and they value benefits such as instant support. The faster claim process powered by AI is a plus, but not at all.

– For insurance agents: While some fear that AI-powered systems might replace human agents, others view AI as a tool that can boost their capability. AI can handle routine tasks and provide agents with collective valuable insights, allowing them to focus on building relationships with clients and strengthening the ability to offer more personalized products and services that meet the needs of customers.

A long journey ahead

Advanced AI algorithms bring a lot of benefits to the insurance industry despite facing challenges in reverse. AI would help to save a lot of time on manual tasks—importing and exporting documents or drafting a brief analysis. However, they cannot make sure the whole complicated process takes place without any mistakes, while insurance relies heavily on accurate regulation and compliance, which is much more prior.

So, at the moment, the traditional approach is preferred: manual reviews, spreading tracking, and periodic audits. Humans cannot be replaced. On the other hand, it seems to be ignorant and cold-hearted to let an AI make a big decision about whether this person gets medical coverage or not.

Yet, once again, for long-term investment in AI, it’s important to seek guidance from legal professionals when building AI agents to optimize operations in the insurance industry; also, a dedicated trainer is needed to leverage employees’ skills.

References: