Actuaries in the age of artificial intelligence
Ben Pring explains how artificial intelligence is likely to change every aspect of the insurance industry and what insurance firms can do to keep up
Actuaries are no strangers to considering odds. Well, what are the odds on this? In recent months, an artificial intelligence (AI) software programme has beaten humans at poker. Libratus – an AI software programme built by a team from Carnegie Mellon University – took down four of the world’s best poker players, in a two-handed game of no-limit Texas Hold’em, and walked away (not literally) with a cool $1.7m.
This, of course, comes hard on the heels of AI’s victory in the game of ‘Go’. Which followed its victory at Jeopardy and chess.
With the current rate of technological advancement, it will soon be inconceivable for a human to beat AI in any such odds-based games of the mind.
Look further afield and AI is winning in lots of other ‘mind-based’ areas as well; in 2015, six of the top eight hedge funds in the US earned around $8 billion based largely – or exclusively – on AI algorithms. In medicine, the ‘new machine’ is quickly surpassing the capabilities of human radiologists. Researchers at Houston Methodist Hospital utilise AI software, which interprets results of breast X-rays 30 times faster than doctors and with 99% accuracy. By contrast, mammograms reviewed by humans result in unnecessary biopsies nearly 20% of the time. In the legal profession, AI-enhanced computer systems are conducting discovery and due diligence far better, faster and cheaper than the most talented team of paralegals in a top law firm. Multiple studies predict that the vast majority of paralegal work can soon be automated. We may reach a point in the not-too-distant future when relying only on humans for discovery might be grounds for malpractice.
The rise of AI is the great story of our time. Decades in the making, the smart machine is leaving the laboratory and, with increasing speed, is infusing itself into many aspects of our lives: our phones, our cars, the planes we fly in, the way we bank and the way we choose what music to listen to.
Within the next few years, AI will be all around us, embedded in many higher-order pursuits. It will educate our children, heal our sick and lower our energy bills. It will catch criminals, increase crop yields, and help us uncover new worlds of augmented and virtual reality. Machines are getting smarter every day and doing more and more; they will soon change our lives and our work in ways that are easy to imagine but hard to predict.
Do you think the actuarial and insurance industries are immune to this tidal wave? Think again. AI will change the insurance business more in the next 10 years than it has changed in the past 200. Lemonade, a relatively new New York-based insurance provider, has already claimed to have processed the world’s first insurance claim exclusively handled by an AI programme.
As such, whether you are managing a large enterprise or just starting your first job, deciding what to do about the new machine – this new cocktail of AI, algorithms, bots and big data – will be the single biggest determinant of your future success.
So, what should you and your company do about the exponential rise of AI? The first, most important but seemingly, for a lot of organisations, difficult step is to stop worrying about what AI might do in the next 20 years and focus on what it will do in the next 20 months. Musing on the potential of machines is the dinner party topic du jour, but while wondering whether machines will act in the best interests of humanity is fun (with a glass of red wine to hand) it doesn’t help you figure out what to do when you get back to the office on Monday morning.
The organisations getting ahead with AI are injecting it into the software behind their products and services, and into the systems that run their operations – today. They recognise that machines are learning to do more and more things and are not fighting it, but embracing it. They understand that ‘systems of intelligence’ are exposing systems (and products, processes, and organisations) that aren’t intelligent. And, lastly, they accept the fact that their customers will, without fail, gravitate to the Google or Amazon price (being generated through prodigious use of AI).
They know what while their competitors are pondering, the moment is right to act.
A great example of what to do when machines do everything is the recent move by H&R Block, which has teamed up with IBM Watson to infuse AI into the spring ritual of submitting one’s tax return. Tax prep is, like insurance, a hugely complex and data-intensive process. Yet, at its core, it is a very routine activity. Unleashing AI on rule and pattern recognition and data crunching has the potential to reduce cycle times and price points and free tax prep staff to spend more time making the annual visit to the H&R office less like a visit to the colonoscopists and more like a financial health check-up, with the corresponding opportunity to sell higher price/margin financial services.
If in private you occasionally catch yourself thinking that the way an insurance claim is made today is pretty much the same as when you first started working, you’re right; technology has only brushed the edges of how the insurance industry works. But in 15 years’ time, thanks to AI, it will be entirely unrecognisable.
What will your company’s products and services look like in 2030? Will they be smart, personalised, full of intelligence and offered at price points that unlock huge new addressable markets? Or will they be only marginally better than they are today, run on systems built in the 20th century, with processes full of paper and duplication and work-arounds?
We are in an amazing time. The emergence of AI is not about simply substituting labour with software; it is about building the new machines that will allow us – you, your organisation – to achieve higher levels of human and corporate performance. When machines do everything, there will still be a lot for you to do. I encourage actuaries to embrace the benefits AI can bring.
Ben Pring is a co-author of What To Do When Machines Do Everything: How to Get Ahead in a World of AI, Automation, Bots, and Big Data (Wiley 2017)