Insurance industry warned of threat posed by fake data

The majority of insurers around the world are failing to ensure the data that feeds their artificial intelligence (AI) systems is accurate, potentially undermining their business decisions.

 

Wednesday 25

After surveying more than 6,300 and IT executives across 25 countries, consultancy firm Accenture found that just a quarter of insurers are validating their data to some extent.

 

This is despite four-fifths believing their organisations must innovate at a more rapid pace to maintain a competitive edge, with insurers increasingly threatened by technological disruption.

 

“Data has always been the lifeblood of the insurance industry, helping carriers make better pricing, risk and operational decisions,” Accenture managing director, Michael Costonis, said. “But insurers must carefully consider how they verify, protect and use this data.

 

“The benefits of accurate, real-time data to consumers could be tremendous – incentivising customers to reduce their exposure to risks and helping them avoid incurring losses in the first place.”

 

The survey found that one-third of insurers have been the target of practices such as bot fraud, spoofed sensor data or falsified location data, while another third think they have probably experienced such an attack but are not sure.

 

Accenture said the importance of validating data could be a key feature to the next wave of industry disruption, with technological change set to “rewrite the rules” over the next decade.

 

Another may be how effectively insurers ‘raise’ AI systems to act responsibly, with the technology predicted to increasingly take crucial business decisions autonomously.

 

This is thought to be particularly important after it was found that more than three-quarters of insurance executives acknowledge that AI is advancing faster than the pace at which their organisation is adopting the technology.

 

Accenture said firms should draw on data science and cyber security capabilities to grade the truth within data, and determine how much risk is acceptable based on the implications of automated decisions.

 

“For all the opportunities data presents, there are real risks if insurers aren’t properly validating their data,” Costonis said.

 

“Thankfully, there are steps insurers can take now, including building a data intelligent practice to inspire confidence in their data grading capabilities and combat the risks from manipulated data.”