impact of big data and ai on insurance - store –impact of big data and ai on insurance nine key...

NCSL Legislative Summit

Impact of Big Data and AI on Insurance

www.pwc.com/analytics

August 2017

Dr. Anand S. Rao, Global AI Lead, Data & Analytics

PwC – Impact of Big Data and AI on Insurance

Contents

2

1. Big Data & Artificial Intelligence Trends

2. Impact on Insurance

3. Risks of Big Data & Artificial Intelligence

Big Data & Artificial Intelligence Trends


Nine key trends are interacting and impacting Big Data, Analytics, and Artificial Intelligence at a speed and scale not observed in the past

4

1 2 3 4Big Data

Big Data is getting ‘bigger’ and ‘faster’

Big Data Technology

Big Data technology is moving from Hadoop to streaming architectures to hybrid ‘translytical’ databases

Democratization

Democratization of data, business intelligence, and data science is accelerating

Open Source Movement

Open source revolution in data, code, citizen data scientist is accelerating access to data and generation of insights

Trend Trend Trend Trend

6 7 8 9Deep Learning

Deep Learning, has been improving in speed, scale, accuracy, sophistication, and scope of problems addressed

MLaaS

Machine Learning, cloud computing, and open source movement are converging to create Machine Learning As A Service (MLaaS) decreasing overall cost of AI

Funding

Big Data funding peaked in 2015; Artificial Intelligence funding, especially machine learning and deep learning are continuing to attract significant investments

Competitive Landscape

The Big Data Landscape continues to grow, while the AI landscape is expanding rapidly with Deep Learning companies growing the fastest across multiple sectors

Trend Trend Trend Trend

5Artificial

Intelligence

Artificial Intelligence is becoming ubiquitous intelligence with the ability to see, hear, speak, smell, feel, understand gestures, interface with your brain, and dream

Trend


Insurance has moved up to the top of the list of ‘most disrupted sectors’

5


Disruptions in Personal Lines Insurance – Usage Based Insurance, Ridesharing, Peer to Peer, Direct-to-consumer, ADAS & Autonomous Cars

6


Disruptions in Commercial Lines Insurance – Direct to Small Business, Drones, Internet-of-things, autonomous vehicles, and emerging risks

7


Disruptions in Life Insurance– Robo-advice, Personalized insurance, Medical

advances, Automated underwriting, decreasing morbidity and mortality risk

PwC – Impact of Big Data and AI on Insurance 9

InsurTech, investments have steadily grown over the past six years with over $17 billion invested

InsurTech Funding and Number of Deals (2010- 2016)

Source: https://www.venturescanner.com/blog/2017/insurance-technology-startup-landscape-trends-and-insights-q1-2017

https://www.venturescanner.com/blog/2017/insurance-technology-startup-landscape-trends-and-insights-q1-2017


The InsurTech market has close to 1,000 start-ups focused in over 14 categories across personal, commercial, life and group benefits sectors


In parallel, Artificial Intelligence funding has been increasing over the past decade with machine learning and deep learning attracting the maximum investments

11

Artificial Intelligence Funding AI Funding by type of Technology

https://venturescannerinsights.wordpress.com/tag/artificial-intelligence/



Artificial Intelligence landscape is expanding rapidly with close to 1500 companies across 13 categories making it difficult for enterprises to track or monitor these companies

12




Deep Learning and Intelligent Agents or Bot companies are the fastest growing areas within AI

13


Big Data is getting ‘bigger’ and ‘faster’

14

Smartphones & Internet of Things will drive both the volume and velocity of data

4.444

180

11

30

80

020406080100

0

50

100

150

200

2013 2020 2025

# o

f S

enso

rs

Dig

ita

l D

ata

Growth of Data driven by IoT

Global Digital Data (zettabytes) No. of Sensors (billions)

Ford GT

• 3000 different signals;

• Handled by six “communication area networks” Generates 300MB of data per second or

• 100GB of data per hour.

http://www.forbes.com/sites/michaelkanellos/2016/03/03/152000-smart-devices-every-minute-in-2025-idc-outlines-the-future-of-smart-things/#7e4407769a71http://www.news.com.au/technology/innovation/motoring/fords-gt-supercar-has-more-computer-power-than-a-fighter-jet/news-story/6e2d23cdeb7a22aa1ae4d53d2075bdcbhttps://smallbiztrends.com/2015/12/60-seconds-on-the-internet.html

1

http://www.forbes.com/sites/michaelkanellos/2016/03/03/152000-smart-devices-every-minute-in-2025-idc-outlines-the-future-of-smart-things/#7e4407769a71

http://www.news.com.au/technology/innovation/motoring/fords-gt-supercar-has-more-computer-power-than-a-fighter-jet/news-story/6e2d23cdeb7a22aa1ae4d53d2075bdcb

https://smallbiztrends.com/2015/12/60-seconds-on-the-internet.html


Artificial Intelligence is becoming ubiquitous intelligence with the ability to see, hear, speak, smell, feel, understand gestures, interface with your brain, and dream

AI that can See AI that can HearAI that can Speak

AI that can Feel AI that can MoveAI that can Smell

5

https://vimeo.com/146492001

https://vimeo.com/146492001

https://www.youtube.com/watch?v=rVlhMGQgDkY

https://www.youtube.com/watch?v=rVlhMGQgDkY

https://www.youtube.com/watch?v=3DWTMh1IdFE

https://www.youtube.com/watch?v=3DWTMh1IdFE

Impact on Insurance


Traditional ecosystem is being enhanced by connected devices offering real-time information that will transform underwriting, loss control, and risk management

Intermediaries

Service Providers

Carriers

Reinsurers

Regulators

Traditional ecosystem

…

Illustrative, not exhaustive

17

IoT and Remote Access

Connected Health and

P4 Medicine

Connected Cars, ADAS, Autonomous

Artificial Intelligence/

Cognitive Computing

The New Normal

Connected Homes

Connected Enterprises


Telematics digitizes ‘ride behavior’ providing insurers with a new way to engage customers, while also supporting the claims process

19

Reduction in crashes

from accident

prevention technologies

which in turn lower

claims costs

20-35%1

Reduced Loss Adjusted

Expenses through 2018

using data from car

sensors

40%2

Savings yield from Automatic ERS,

ACN, and hence FNOL initiation

5-15%1

Savings realized from telematics data

enabled technologies to reduce fraud and

theft

30-80%1

Driver Needs

Using Google Maps to

navigate their trip using the fastest route

Sources: 1. Cisco IBSG estimates, 2011; 2. PwC research and analysis, Celent forecast


As seen in the broader FinTech space, InsurTechs get traction and evolve their model to tap into existing profit pools

20

Metromile is founded to enter the Insurance market with a customer centric Pay-Per-Mile value proposition.

Metromile partners with Uber to provide coverage when the driver is not covered by Uber commercial policy.

Metromile acquires Mosaic Insurance… with the potential to underwrite its own policies in 50 states.

Also raises $191.5M that will support its expansion plans.

Mark Cuban one of the main investors

2011 2015 2016 Operating as a MGA MGA to become a Carrier


The Climate Corporation revolutionized yield management and crop insurance for growers using machine learning techniques to model yield by individual fields and crop

Source: http://www.climate.com/; FastCompany, October 7, 2013

21

http://www.climate.com/


Discovery has demonstrated a way to break the ‘vicious cycle’ and deliver personalized underwriting by using technology, information and analytics

Underwriting based on Vitality Age –Discovery South Africa

Source: www.discovery.co.za

22

Risks of Big Data and Artificial Intelligence


Risks & Challenges of AI

“I am in the camp that is concerned about super intelligence. First the machines will do a lot of jobs for us and not be super intelligent. That should be positive if we manage it well. A few decades after that, though, the intelligence is strong enough to be a concern.” Bill Gates, 2015

24

“I’m increasingly inclined to think that there should be some regulatory oversight, maybe at the national and international level, just to make sure that we don’t do something very foolish.” - Elon Musk, 2014


Emerging research and theory aims to fill these gaps and build trust by ensuring understandability and societal benefit are priorities of emerging AI

26

Used in sentencing, AI programs trained to predict recidivism trained on facial features of criminals appear to be racist1

Tay, a Microsoft chatbot, released to interact with the public began tweeting racist and inflammatory remarks in under 24 hours2

If an autonomous vehicle must choose between potentially injuring it’s passenger or a pedestrian, it is unclear how (or why) the algorithm choose3

Many advancements in AI increase automation, often resulting in job loss and replacement4

Bias-Prone Difficult to ExplainWith No Sense of

MoralityWith Societal

Disregard

Both arrested for stealing


Key aspects of AI are raising concerns with respect to Trust & Transparency resulting in multiple efforts around Explainable AI and Beneficial AI

Beneficial Artificial Intelligence (BAI)

Developing AI to with a consideration to maximum societal benefit in applications for justice, economy, and social policy

Ethics

ValuesFuture AI

Development

Diversity

Explainable Artificial Intelligence (XAI)

Building AI models with accountability, and the ability to describe or depict why a certain decision was made by the algorithm

Image Source: Gunning, DARPA I/2O, 2017

PwC Data & Analytics

What are the risks of AI?

28

Bias-Prone Difficult to Explain

No Sense of Morality Societal Disregard

• Biased dataset• Biased developers

• Black-box approach• Accuracy vs

Explanation trade-off

• No ‘gut-check’• No values

• Disregard to jobs• Governance of

decisions

“

We cannot blithely assume

that a superintelligence

will necessarily share any

of the final values

stereotypically associated

with wisdom and

intellectual development

in humans

NICK BOSTROM IN THE SUPER-

INTELLIGENT WILL


What are the considerations for policy makers?

29

“

“where regulatory

responses to the addition

of AI threaten to increase

the cost of compliance, or

slow the development or

adoption of beneficial

innovations, policymakers

should consider how those

responses could be

adjusted…”

PREPARING FOR THE FUTURE OF

ARTIFICIAL INTELLIGENCE, OFFICE OF

THE PRESIDENT OF USA, OCT 2016

02Enterprises

03Society

Policy Implications

for Responsible AI

Individuals• Privacy & Confidentiality• Personalization• Explanations

04Government

01

• Trust & Transparency• Responsible &

Explainable AI• AI Governance

• Automation & universal basic income

• Education & Skills• Regulation – Monopolies,

Consumer protection

• Bias & Discrimination• Diversity &

Transparency• Beneficial AI

PwC

© 2017 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see www.pwc.com/structure for further details. This content is general information purposes only, and should not be used as a substitute for consultation with professional advisors.

Dr. Anand S. Rao

Global AI Lead, Innovation Lead


+1 617-633-8354

[email protected]

Twitter: @AnandSRao

Thank you!

30

mailto:[email protected]

impact of big data and ai on insurance - store –impact of big data and ai on insurance nine key...

Documents