Insurers' caution could pay off in getting AI right

Like many of my colleagues, I’m tired of the stereotypes that continue to plague the insurance industry – that it’s unexciting, nerdy, slow to adopt technology advancements, unable to acquire good IT talent and all the rest. This is beginning to change with insurtech advancements, but the industry can’t seem to shake the “technology laggard” label.

There now appears to be one reason to actually slow insurers’ rush to technology nirvana, and it has to do with the spawn of analytics: artificial intelligence. The good news is that insurers of all sizes have realized the promise of analytics, and have applied that discipline to everything from customer experience excellence to improved loss control measures. But the advent of AI carries with it a huge investment in time, effort and dollars, and most often involves the acquisition of outside (insurtech) support. The fact is that many insurers that have made that investment have yet to prove its merit, either because it’s too early to realize results, or they don’t understand how to leverage its full potential.

Consider that insurtech funding totaled $2.3 billion in 2017, a 36% increase over 2016, according to Willis Towers Watson. Approximately 83% of insurtech ventures comprise an insurer or reinsurer as the investor, while 17% of ventures have no such investors, notes the report. The firm places insurtech ventures into four categories: product and distribution (the largest category); business process enhancement; data and analytics; and claims management.

There are 70 companies working on data and analytics, focused on “analyzing new sources of open source or other external data and/or extracting insights to enable data-driven decision-making,” the report said.

You can’t have AI without data and analytics, so you can see where this is going: What insurers are investing in today is likely not true AI, but rather just its underpinnings.

Kiyoshi Ota/Bloomberg

The enthusiasm related to AI clouds our ability to understand its reality. The Economist reports that in the last quarter of 2017, public companies across the world mentioned AI and machine learning in their earnings reports more than 700 times, seven times as often as in the same period in 2015. The Economist quotes Tom Seibel, a Silicon Valley veteran, who states that there are so many firms peddling AI capabilities of unproven value that someone should start “an AI fake news” channel. Even a large vendor in the insurance data analytics space reports that it has had to hold special meetings with its employees to explain the reality of AI’s promise in order to temper customer expectations.

As you can see by this Twitter post, there is still much confusion about what AI is and what it does.

So, what is the reality of AI in insurance? Many blindly accept the Google definition (the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages).

Michael I. Jordan, professor in the Department of Electrical Engineering and Computer Sciences and the Department of Statistics at UC Berkeley, believes we are not there yet. In a recent Medium post, Jordan notes that the current focus on doing AI research via the gathering of data, the deployment of “deep learning” infrastructure, and the demonstration of systems that mimic certain narrowly-defined human skills — with little in the way of emerging explanatory principles — tends to deflect attention from major open problems in classical AI.

“These problems include the need to bring meaning and reasoning into systems that perform natural language processing, the need to infer and represent causality, the need to develop computationally-tractable representations of uncertainty and the need to develop systems that formulate and pursue long-term goals. These are classical goals in human-imitative AI, but in the current hubbub over the “AI revolution,” it is easy to forget that they are not yet solved,” he says.

On this most experts agree: investing in AI is not a simple matter of pulling a rabbit out of a hat. Insurers making large investments in AI and insurtech support may face the slings and arrows of trying to decipher how to garner its greatest and fastest return, but insurers that hesitate to invest may also find themselves vulnerable to the competition that got it right and is now leading the charge.

To get it right the first time, maybe being a laggard isn’t such a bad thing.