AI growth offers insurers opportunities
Artificial intelligence is not about robots taking over the world, but about leveraging the powerful technology to capture and analyse vast quantities of data, and making the best use of that information.
That's the belief of Noel Pearman, Bermuda professional lines underwriter and cyber product lead at XL Catlin, who has been speaking on the subject in the US.
For the insurance sector, one likely manifestation of AI will see customers purchase increasingly tailored policies based on data directly relevant to the way they live their lives.
Mr Pearman explained: “One of the things we are doing now is capturing data at a rate that we were never able to do before.”
The prevalence of “The Internet of Things” devices that collect real-time data will allow insurance companies to mine information and offer better rates to customers whose behaviours represent a lower risk, such as those with better driving habits.
“Multiple sensors on vehicles that have telematics are now capturing details such as how fast someone is driving, how fast they are braking, how often are they having close calls,” said Mr Pearman.
The data is being collected at a rapid rate, and AI will help to process it.
“Now we are building systems where we can actually take a look at all this data and make intelligent decisions and inferences on it. That's the type of thing that AI is able to do,” said Mr Pearman.
“Computers are able to process better data and differentiate between insurers — we're starting to be able to do that for the first time.”
At the start of the month, he was at the Insurance Coverage Litigation Committee CLE Seminar in Tucson, Arizona, where he spoke on a panel that discussed AI and its impact on the insurance industry. Lawyers were also on the panel, and one branch of the discussion considered liabilities, such as who can be sued if a decision is made by AI.
“We are interfacing with machines that are learning, getting better at their jobs and over time can create liabilities for companies that rely on those decisions.
“It sounds muddy, because it is at this time. It is all aspirational. That panel was thinking through how does this play out in the legal realm.”
Mr Pearman said XL Catlin views AI as an opportunity to provide “value up and down the insurance chain”. He said the company and others like it are working in the space by investing in AI startups, providing insights and access to data, or providing underwriting capacity.
An example for XL is Oxbotica — a British autonomous vehicle software company.
“It is a car company using AI to help figure out a self-driving vehicle. It might seem a strange partner for an insurance company, but it's an excellent way to get a ground up view of the risks and challenges of self-driving cars.”
Mr Pearman acknowledged that there are widely differing understandings of what AI is, and what it is capable of achieving. He mentioned AlphaZero, an AI developed by a Google-associated company, that achieved “superhuman” abilities at chess within 24 hours of encountering the game and its rules.
“It was like an alien playing chess. It was making sacrifices and moves that no one had contemplated before, and it was winning in ways that no one had thought possible. This was achieved in a day. But remember, chess is a very narrow application and has perfect information about what you and your opponent have.”
In the real world, and in insurance, there are many nuanced and grey areas, but Mr Pearman sees change coming.
“The way AI is working in practice, it will eventually allow people not to do the ‘legwork things' and spend more time on the higher-value things. So it will change the landscape.
“Figuring out what the effect will be on the job landscape is difficult. But what I can foresee is that we have skill-sets beyond what we can get machines to do.”
An example is unstructured data, such as when pertinent information needs to be identified and gathered from sentences, paragraphs and longer text documents rather than being found in “nice Excel spreadsheets”.
Asked about concerns people might have that AI is an existential threat to mankind, a view aired by among others SpaceX entrepreneur Elon Musk, Mr Pearman said: “What it isn't right now is robots taking over the world. What it is, is very narrowly focused AI. It can do one thing, probably exceedingly well, but only that one thing.”
However, he said the concerns were legitimate but that would be getting into “general AI” rather than the narrow AI that exists today. However, he added: “These things have to be done intelligently.”
Bermuda was also represented at the Tucson event by Sandra DeSilva, managing director and chief software architect of Bermudian-based software engineering company Nova.
Mr Pearman, who is in the running for the cyber-risk industry person of the year title from Advisen, said it was good for Bermuda to be on the world stage at events such as the conference in Arizona, and to be involved in discussions on evolving technologies.
He is also positive on the blockchain and technology incubator ambitions of the Bermuda Government, referring to them as being “right along the lines that Bermuda should be thinking — future focused”.
Noting Estonia's huge investment in digitising its economy, he said Bermuda was almost the perfect environment for such a shift.
“They have 1.3 million people; at that scale they ran into some issues. But what is 60,000 people in a wealthy country? We are quickly updating our technical capability, our internet capability and connection capability. We are almost a perfect place to start understanding technologies like this and roll things out.
“Sixty thousand people is so manageable. It makes a lot of sense to pioneer stuff like this; to use intellectual capital to set ourselves apart and make a difference. Given the appropriate amount of care and attention, things like this are the way Bermuda needs to go.”