With more pressure than ever to offer competitive pricing, insurers are seeking innovative ways to leverage additional data sources in underwriting. Today, there are over 4.62 billion social media users globally, leaving endless amounts of personal data across social media platforms like Facebook, Twitter, Instagram and LinkedIn.
The need for more data sources to augment and support internal processes is growing, especially with the rise of predictive analytics, AI and new demands to improve the customer experience.
Social media data provides insurers with an opportunity to gain insights into a customer's risk exposure in real time. But it comes with many challenges. It can be used—but take care!
Some insurers have already jumped on the bandwagon
One auto insurer created a personality-type assessment based on certain choices and actions of potential clients. These included what athletes the client likes/follows, how concise their writing was and how often they used exclamation marks in social media interactions. The data enabled the insurer to determine if the client was overconfident or reckless, traits associated with many high-risk drivers.
A life and health insurer in the U.S. tested behavioral data gathered from online retail sites and third-party databases as inputs for predictive modeling to determine the health risks of over 60,000 applicants. Examining user behavior helped the insurer get results similar to traditional medical examinations. Each application cost the insurer only $5, instead of $250 to $1,000 for conventional medical tests, according to a Deloitte report.
AI to the rescue
Scalability is a critical challenge with social media usage in underwriting. Manually collecting vast quantities of social media data is impossible. However, insurers can partner with third-party data vendors that use AI to scrape data from users' profiles.
AI tools, equipped with machine-learning algorithms, can collect written and visual consumer data and build predictive models faster than human agents. Predictive models let underwriters gain a more detailed assessment of an insured’s level of risk.
Sentiment analysis, equipped with natural language processing (NLP), is a machine-learning technique that analyzes and interprets text. Sentiment analysis can take in written user information at scale and use it to assess a client's behavior. For example, sentiment analysis can read, analyze and collect information from a business's review section, flagging any potential risks that require further investigation.
Machine vision uses software algorithms to assess images based on existing data sets already evaluated by humans. Insurers can leverage machine-vision applications to investigate the photos and videos to discover more about a client's lifestyle, including eating, exercise and smoking habits.
Legal and regulatory pitfalls
Experts say insurers using social media data to fight fraud are on solid legal ground. But, what about underwriting? Regulation is evolving. Proceed with caution.
Data mining through social media may violate privacy laws, such as the E.U.'s General Data Protection Regulation (GDPR).
Closer to home, the New York State Department of Financial Services says they're concerned insurers could use their own algorithmic underwriting systems to discriminate against consumers illegally.
New York’s insurance laws, and similar laws elsewhere, prohibit the use of race, national origin, lawful travel, mental or physical disabilities or traumatic experiences such as domestic abuse in any aspect of insurance underwriting.
Insurers must ensure that the external data sources they use meet their antidiscrimination requirements. Additionally, carriers must be transparent to customers about the content of the external data and its source when being used to increase premiums or deny a customer coverage.
One expert says insurers can test their algorithms for discrimination by examining the algorithm's results. For instance, insurers can test their algorithms by omitting standard customer data and only using information about the client's race. If the algorithm can predict the customer's premium by only using race, the model is too dependent on protected personal attributes.
See also: Personal Connections Via Social Media
Does using social media data violate trust?
Collecting clients' social media information without consent or transparency can feel like a violation of trust. If not done sensitively, this can severely damage an insurance company's reputation, brand and relationships with policyholders. Always ask prospects and customers for permission to view their social media posts.
Lawyer Tyler Dillard compares the use of social media data to previous decisions arrived at on the issue of genetic testing for insurance. For example, while AI may suggest certain typing habits (e.g., excessive use of exclamation points) are correlated with bad driving, the causal link is dubious and supporting evidence is sketchy, unlike genetic testing. Insurers that use AI for big data analysis of social media should emphasize explainability above all else: Why did our program draw this connection?
From a proportionality perspective, analyzing social media data in such a granular way may only produce marginal benefits at the cost of severe regulatory consequences and offending customers.
Of course, the accuracy of the source data on social media is suspect. The online personas users construct are not always accurate reflections of reality. Moreover, insureds can seriously undermine the underwriting process by learning what information carriers are looking for and publishing content they believe will help them get a lower premium.
Despite its challenges, social media remains a great tool for enhanced customer intelligence and developing more personalized policies.
Although social media data offers a trove of growth opportunities for insurers, the legal challenges, brand-affinity risks and lack of data verification are causing many to steer clear of this underwriting method. In fact, out of 160 insurers investigated by New York state, only one used social media for underwriting.
To mitigate risks, insurers should focus on rewarding “good” behavior on social media rather than punishing “bad” behavior. In other words, offer discounts to desirable risks. Many regulatory bodies are slow to respond with clear guidelines on big-data analysis of social media. Err on the side of caution.
As social media data usage in underwriting is adopted, insurers must be transparent with customers about using and obtaining the data. This step is imperative, as undisclosed use of client social media data can infringe on privacy laws and antidiscrimination laws and leave customers feeling violated.
Leveraging consumer external data remains critical for insurers today. Insurers that comply with regulatory requirements and emphasize explainability and proportionality can mine magnitudes of external customer data and turn it into actionable information. Of course, they’ll need modern rating and underwriting systems to leverage that information.