MIT Researched 16,625 Papers on Artificial Intelligence — Here’s Why Insurers Should Care

MIT researchers conclude that machine learning has the potential to create great efficiencies; that goes for its use in insurance too.
By: | February 6, 2019

The Gist: To learn where artificial intelligence (AI) is headed in the future, MIT Technology Review took a look into the past. Researchers analyzed 16,625 papers from 1993 to 2018 to learn how the technology, and the research behind it, has evolved.


MIT reveals that AI research trends popularize then fade away over the years — and new paradigm shifts come at approximately every decade. Sometimes they’re brand-new technologies. Sometimes existing theories are resurrected. In any case, we’re due for yet another shift in AI research in the 2020s.

What does AI look like now? AI went through a major seachange recently, going from knowledge-based systems to machine learning. Knowledge-based systems are algorithms based on rules that can give machines common sense and human knowledge. With machine learning, systems teach themselves the rules.

As researchers studied knowledge-based systems they found a major problem: “there were simply too many rules that needed to be encoded for a system to do anything useful. This jacked up costs and significantly slowed ongoing efforts,” the report said. “Machine learning became an answer to that problem. Instead of requiring people to manually encode hundreds of thousands of rules, this approach programs machines to extract those rules automatically from a pile of data. Just like that, the field abandoned knowledge-based systems and turned to refining machine learning.”

What’s Next for AI?: That’s the billion-dollar question. MIT stopped short of predicting what the next decade of research will look like. Perhaps an older technique will regain favor or an entirely new technology will emerge.

Pedro Domingos, a professor of computer science at the University of Washington and author of The Master Algorithm offered his thoughts:

“The 2020s should be no different,” he said, “meaning the era of deep learning may soon come to an end. But characteristically, the research community has competing ideas about what will come next—whether an older technique will regain favor or whether the field will create an entirely new paradigm.”

What’s Does AI Have to do With Insurance?: Plenty. AI will fundamentally change the way we process claims, assess damage and underwrite risks in the near future. In an article titled Here’s How Artificial Intelligence Is Poised to Transform Insurance Underwriting for the Better, we took our own deep dive into AI. We found AI will soon deliver more accurate risk assessments, superior customer experiences and serious cost benefits for insurers.


“Imagine workers’ comp underwriters quickly analyzing thousands of pages of medical bills and health records to predict injury risks,” the article said. “Imagine commercial property underwriters seamlessly incorporating external data from city governments, regulatory agencies and news sources to get a clearer picture of a risk. Imagine a shipping company that’s charged more for premiums when sailing into bad weather or hostile territory.”

Further Reading (and Viewing): AI’s reach will be far and wide, affecting almost every part of the risk management landscape. One of the coolest applications is its use by first responders. Futurism explains how police, fire and others are using AI to deliver critical aid in the wake of natural disasters.

We also recommend this fascinating Ted Talk from Domingos explaining the next 100 years of your life. He offered this enlightening quote: “a decade’s worth of progress now is roughly equivalent to the entire 18th Century, or to the entire first millennium AD.” and said that the possibility of living forever is closer than you think.

Jared Shelly is a journalist based in Philadelphia. He can be reached at [email protected]

More from Risk & Insurance

More from Risk & Insurance