Currently showing: Other


24 Feb 15 16:38

Last year when I attended the World Economic Forum in Davos, I was struck by the number of people talking about data. This year, the conversation has shifted...a lot! Most people I came across at the WEF and in the weeks following are now talking about cognitive computing. People readily accept now that we can mine large data sets of structured and unstructured data for new insights. This is no longer a debate about false correlations and getting the questions right (although still an important consideration).

The discussion has shifted to the power of cognitive computing, and the ability of the "machine" to think through problems it could only "search on" before. Things like IBM's Watson has changed the way we think about using computers. And this is only the beginning.

The rise of cognitive computing will have massive impacts on the business world as we know it. Jobs will change materially over the next decade. Some skill sets will no longer be needed. The debate is around when and to what degree, not whether it is possible.

Autonomous cars, 3D printing, drones, Artificial Intelligence and robotics will all impact our daily lives. But what about insurance? What impacts will we feel? Which ones will we embrace and drive, and which will we resist? And most importantly, can we adapt fast enough?


Category: Other


7 Comments

Krishna Burli - 25 Feb 2015, 6:16 a.m.

In Insurance, "Beyond Data...." requirements are many...between PROMISE & FULFILLMENT.
Ambiguous, understanding the way in which the promise has been made, the promise with hidden intentions & agendas so to say......it goes on & on..

Lack of deliveries on the basis of promise, legal recourse to redress of promises unfulfilled as per contract, is another big area of litigation in Insurance.
Gaps between Promise & deliveries due to processes, documentation, accounting which are commensurable with the RISK COVERED subject to policy terms & conditions, need to be properly UNDERSTOOD BY THE CLIENT the way it is promised. Big data comes to the rescue of this DOCUMENTATION part of contracts & legalities.

Cognitive computing. is a step in the right direction, as cognition of GAP is more required in the agreed structures, be it from CLIENT angle or INSURER's angle.

"What impacts will we feel? Which ones will we embrace and drive, and which will we resist?"...The answer to this question can be simple....The path of least resistance to Client & Insurer with Big Data results authentic & agreeable

"And most importantly, can we adapt fast enough?"
Adapting is at different perspective to Client & Insurers, based on the speed economy & authenticity....initial resistance may curb the speed but, however, the ease of BUSINESS DOING in a convenient, quick & agreeable manner can boost the confidence & impact of GOOD FEEL con creep in to enhance the speed of "cognitive computing"
The above views are of my individual, with 36+ yrs of Non-life Insurance Domain expertise with Risk Management & I am subject to corrections in case I am convinced otherwise.

Francisco Jimenez - 26 Feb 2015, 5:17 p.m.

This article in the LA Times yesterday details the Deep-Q Network computer, developed by Google's DeepMind project, which can mimic the way human brains learn from new experience. It was tested by playing 49 classic Atari games scoring "at or better than human level" on 29 of them.

What's really cool is it even found a few tricks unknown to the programmers in some of the games. It has a long way to go, but this is certainly a big step in the AI direction.

http://www.latimes.com/science/sciencenow/la-sci-sn-computer-learning-space-invaders-20150224-story.html

Paritosh - 27 Feb 2015, 2:25 p.m.

Let's kill the consideration over correlation and getting the question right, a little further on please :o)

Correlations, often get confused with causation, which is not actually the case (e.g. http://goo.gl/SGJOCN).

So, what is required is to confirm via scientific (valid) methods, to prove whether a correlation is worthy or not. If a human mind can think it's worth (beating the time spent on validation), then there is the money! Perhaps the companies that does it sooner than than others, would reap loads of benefits, as is evident from the pharmaceutical industry.

Autonomous cars, 3D printing & drones, do present plenty of risk to underwrite about.

As far as AI is concerned, I guess Insurance industry should be very scared of it, cause it solely aims at reducing the risks caused by human interference.

Alicia Montoya - 12 Mar 2015, 6:47 a.m.

If you want to know what the future of tech has in store, South by Southwest is the place to be!

"This year top speakers will tackle how to bring ideas that might seem like science fiction to the real world. Google's Astro Teller, who heads the tech giant's X division, will discuss making some of Google's most ambitious projects, like self-driving cars or balloon-powered Internet, a reality. United Therapeutics CEO Martine Rothblatt will talk about tech innovations intersecting with artificial intelligence, immorality and the future of self. The Winklevoss twins, Tyler and Cameron, will be on hand to discuss what Bitcoin needs to do to succeed as a mainstream currency."

And that's before mentioning all the up-and-coming talent that the conference is known for. Follow it on twitter using #SXSW

http://bigstory.ap.org/article/a9de758713bf4cada393b2e453663311/flying-cars-robot-petting-zoo-its-south-southwest-time#

Alan Brasunas - 18 Mar 2015, 2:17 a.m.

Technology, big data, and cognitive computing all will continue to impact the insurance industry, largely for the good, let’s hope. The promise of reduced insurance premium when evidenced or suggested by good behavior, which serves to reduce overall losses, serves as a compelling inducement to the buyer of insurance.

If you haven’t read the article recently appearing in The Economist titled “Risk and Reward,” I highly recommend it. The URL: http://www.economist.com/news/finance-and-economics/21646260-data-and-technology-are-starting-up-end-insurance-business-risk-and-reward?fsrc=scn/ln_ec/risk_and_reward

Jeffrey Gonlin - 18 Mar 2015, 7:27 a.m.

Could underwriters be replaced by machines? I agree with Jayne – “…the debate is around when and to what degree, not whether it is possible.” Consider the Oxford study on occupations that are likely to be threatened by advances in IT. Where did Insurance Underwriting rank? See for yourself: http://www.oxfordmartin.ox.ac.uk/downloads/academic/The_Future_of_Employment.pdf.
Data is just the raw material – and it’s increasingly becoming available in heaps. But to compliment Big Data we need equally Big Analysis. Machine learning will change a lot. It’s largely about pattern recognition, after all, and machines are learning to sort through much faster (and eventually better) than humans. This will be a profound enabler. Isn’t it already happening on Property Cat? On Personal Lines?
Some have drawn a contrast between Property and Casualty by asserting that Casualty events, because they are human-made, defy categorization and cannot be modeled. I disagree. Humans in the aggregate are not capricious; human events do unfold within parameters. Consider them confidence intervals of sorts.
Sorry to all those business majors in the business, but what a great time to be a Sociologist! One upside from Big Data: in the next decade we will learn more about social phenomenon in fact than we learned from the past 100 years of theorizing. And this learning will be exponential. The challenge – and the opportunity – is to learn how to apply it. Machine learning, coupled with a profoundly anthropological approach to insurance, will be the way forward.
Consider the pace of change and the potential scale of risks from emerging issues. Do we have the luxury of waiting for consensus on the question of causation, final adjudication and eventual collection of statistics before we can transfer these risks for a price? Reality keeps moving forward! If we can do better, we have an obligation to do so - an obligation to protect capital from adverse outcomes, but also to craft solutions to protect businesses and society.
Typically, insurance relies on the past. But here, that won’t do. One must anticipate the future. The good underwriter is always looking ahead, anticipating. But do we even know what to monitor? If we can scan more systematically, more comprehensively by letting machines scan for us, we’d be fools not to.
By gazing at the stars night after night humans learned to navigate the globe. Big data + machine learning will be like gazing through the Hubble telescope. New structures will become visible, new theories will emerge, deeper understanding will result. It will happen. The question is, as Jayne put it, can we adapt fast enough? I believe making our current model obsolete and replacing it with something better would be the greatest contribution to the industry and to society that we could possibly make. We have to see the vision and we have to want it…

Alicia Montoya - 6 Jun 2015, 8:48 p.m.

Just came across an article that beautifully explains how software bots are supporting white collar workers' daily tasks, enabling the latter to focus on value added work.

The example the author uses is a hedge fund analyst but the same could apply to actuaries and underwriters. Having just spent a few weeks looking at the kind of data work they do, copying and pasting data into different programs and formats to then "do their magic" as I'm calling it (and yes, underwriting is an art!), it seems like a 'no brainer' to want this software support. That way, as the example goes, the underwriters and actuaries would have saved a morning of data manipulations, which they would have spent focusing on the client's needs, developing and delivering products and services to add value and support their client's growth. Bring on the bots, I say!

http://www.zdnet.com/article/white-collar-automation-will-bring-new-industrial-revolution-says-ceo/


If you would like to leave a comment, please, log in.