Advantages and disadvantages of artificial intelli

  • Detail

Advantages and disadvantages of artificial intelligence used by insurance companies

in the near future, when you wake up, your watch records how long you slept, and monitors your heartbeat and breathing; When you drive to work, the car sensor will record your speed and braking; You buy some breakfast on the way, pay electronically, and then the transaction information and the calorie content of the food are recorded

then unfortunately, you have a car accident. When you call the insurance company, you will be answered immediately. The operator at the other end knows your name and kindly talks with you about your pet cat and your favorite football team's performance on the weekend

you are chatting with a chat robot. The reason it "knows" you is because insurance companies are using AI to gather information about you from social media. It also knows more because you have agreed to let it monitor your personal devices in exchange for cheaper insurance premiums

this is not science fiction. More than three-quarters of insurance company executives believe that AI will completely change the insurance industry in a few years. According to McKinsey futurists, by 2030, AI will mean that your car and life insurance costs may change according to which route you decide to take

it will sell to you with the promises of more personalized service, faster claim processing speed and lower insurance cost, and it will fulfill these promises in most cases

but there are also moral risks - data privacy and discrimination. The insurance company may use your data to calculate how much insurance premium you are willing to pay. It may also sell information to third parties. In addition, AI may determine your risk because of your age, gender, income or race


although the insurance industry is usually notorious for taking customers' money and refusing to pay claims, it is also a very competitive industry. If competitors use AI to keep profits while reducing premiums, companies that react slowly may not be able to survive

in order to provide a lower premium, insurance companies need to know that the actual risk of individuals is low. Supporting technology is IOT, which is the general name of billions of connected sensors embedded in various objects we use every day. Among them, watches, cars, fitness trackers, home assistants and many other things, they together constitute the "ecosystem with long service life" of sensors

over time, the data collected by IOT sensors allows insurance companies to develop personalized risk profiles based on individuals' actual behavior, which is called behavioral policy pricing

get "smart"

in order to reduce your housing and property insurance costs, the insurance company will access the AI center running your smart home through its IOT sensor ecosystem

if there is a burglary nearby, the smart home center will know because it is connected to the insurance company. Then automatically detect whether the lock and alarm operate normally or start them, and dial the alarm immediately when there is a problem. In order to control the fire risk, IOT sensors will monitor heat, humidity and smoke. If the gas stove is always on, the smart home center will turn it off before it becomes a problem

in order to calculate a lower car insurance premium, the insurance company may want to monitor the way you drive and maintain your car

to get a more favorable health insurance rate, the insurance company may need to check your medical records and require you to wear a fitness tracker

a new service industry will appear. Professional companies that deploy IOT sensors and collect data will cooperate with insurance companies to form a new business ecosystem. The whole insurance industry will shift from pure passive insurance to active and risk minimization insurance

this sounds optimistic. However, in the narrow pursuit of minimizing the risk of joint venture insurance in the field of material experiment, there are also broader risks


a very obvious risk is characteristic analysis - because you belong to a specific population group, you are judged as a higher or lower insurance risk

AI can now distinguish risks into hundreds of factors, and the algorithm scans these factors to identify previously unidentified risk groups

but these conclusions may inadvertently produce discrimination. There have been many examples showing that AI algorithms inadvertently magnify the impression of bias

the criminal prediction case in Durham, England, illustrates this problem. The fatigue testing machine there is also divided into different types. The police use an algorithm that can better predict whether the accused should be granted bail. What it does is discriminate against the poorer people according to where they live

opportunity pricing

in addition, there may be more personalized discrimination

it is well known that genetic discrimination is a problem - the health or life insurance company reveals your genetic tendency to certain diseases based on your DNA, and then the insurance company asks you to increase the premium or even refuse to underwrite

AI opens a new field of personalized discrimination based on the information it can collect from your behaviors and preferences

first of all, AI may provide a lot of data to tell insurance companies about your consumption habits. Where do you shop? What do you want? When did you buy it? Are you looking for bargains or full price

knowing all this will help the insurance company evaluate whether it can charge you the highest premium

industry insiders believe that this is only the way of market operation, but it has become a very questionable practice when obtaining personal information is unprecedented at present

loss of privacy

insurance companies may also use data for purposes other than risk assessment. Given the value of data, insurance companies can sell it to third parties for various purposes to offset the cost of collecting data. Advertisers, marketers and the government are all eager for detailed demographic data

contrary to what people think, these data are not the property of their related people, but belong to the person who pays. Consumers must be protected by law and must not use their data for other purposes

manage risks

any powerful new technology has advantages and risks. We should clarify the benefits and control the risks at an acceptable level from the perspective of customers. Of course, the risk of insurance companies managing risk is very ironic

insurance companies have a job to do to ensure that customers believe that the benefits of AI far outweigh the disadvantages. They need to adopt transparent, fair, if not benevolent, practices that will help to achieve greater interests, and they must not only consider profits

Copyright © 2011 JIN SHI