From attitudes toward life to one's reading history or purchasing behavior, AI is used to analyze large amounts of personal data. Systems that create scores based on an individual's potential abilities, personality and credibility are used even in Japan.
Yet what kind of society would emerge if these scores become widely used and individuals' lives become more quantifiable? Will it be an objective and fair society or an oppressive society in which we are monitored through scores that pry open our inner worlds?
"Tell us about yourself."
A website for J.Score Co., a consumer finance unit jointly owned by Mizuho Bank Ltd. and Softbank Corp., has 150 questions that are scored by AI out of a maximum 1,000 points. Those who achieve higher score receive higher upper limits and lower interest rates on loans.
Questions include "Does it bother you when shop staff get your order wrong?" and "How often do you listen to the radio?" while also inquiring about what newspapers you read, your life experiences, exercise habits and other matters. Many questions seemingly have no relationship to one's ability to repay a loan.
However, AI works with large amounts of data and uses "profiling" to identify relationships. It can infer information a person has not revealed and make predictions about future behavior and their chances of committing a crime, among other outcomes. At times, AI can even deduce information that individuals themselves are not aware of, such as whether they purchase large quantities of cosmetics or have a tendency toward depression.
"We want to collect a lot of answers so the AI learns and improves its score accuracy," J.Score President Ryuichiro Omori explained. Since the service was launched in September 2017, over 500,000 people have completed the questionnaire.
"We expect to surpass 1 million this year," Omori said proudly.
Currently, the scoring system is only used to establish credit limits, but in the near future, the company is hoping to launch an information bank (see clip) and start selling the scores to other companies. J.Score has already received many inquiries from rideshare services and other sharing economy operators like Airbnb.
Other companies are pursuing similar initiatives. Motivated by China, which has a head start in scoring behavior, companies such as Yahoo Japan Corp., NTT Docomo Inc., and LINE Corp. have announced their entry into the scoring business. We can expect big data -- histories of purchases made with cashless payments, location information, online browsing data and other metrics -- owned by each of these companies to be fed into their scores.
With more companies such as Airbnb facilitating business among strangers, the demand for such scores is increasing. This system may actually work more favorably than existing indicators for young people with meager finances or freelancers not affiliated with an organization.
But what would happen if these scores provided to companies were to become universal scores used widely throughout society?
"If we were to find ourselves receiving unfavorable treatment or not being able to use hotels or restaurants without having a score, many would end up providing personal information to raise their scores. This would result in scores becoming more accurate and their impact becoming even wider, and we would find ourselves heading toward a score-based society at an increasingly rapid pace," Tatsuhiko Yamamoto, a Keio University professor knowledgeable about constitutional matters, said.
Concerns are not limited to privacy, as the issue pertains to the broader transformation of society. J.Score, for example, raises a person's score if they take actions recommended by the company. Their score goes up when they read the summaries of dozens of books prepared by the company, or when their smartphone measurement for the number of steps they walk exceeds a target value.
Yamamoto is apprehensive, saying: "If people start thinking and behaving in a way that conforms with good behavior as defined by AI, I think we risk losing social diversity."
"To limit the influence that scores have on our lives, we should prohibit the use of scores in areas that deeply impact our lives, such as employment, marriage, university applications and so on," said Ryoji Mori, a lawyer who serves on a government commission investigating the accreditation systems for information banks.
Mori believes the information used for analysis needs to be limited. For example, data obtained from sensors attached to IoT devices includes physical information such as heart rate and one's location, which enables inference of subjects' mental or physical health.
Similarly, the reading history and subscription information that J.Score obtains from users allows it to infer the beliefs and convictions of particular individuals. The company explains that "there is no problem, as users consent to answering the questions."
But Mori points out that "as the scores come to have greater influence, users will have no option but to provide information, and the concept of consent will have no substance." Mori believes that when it comes to sensitive information such as beliefs and one's ideology, guidelines are needed such as those used in the financial industry that protect personal information, prohibiting the acquisition of such information even with someone's consent.
What kind of society would scores create? Joji Shishido, a University of Tokyo professor who specializes in the Constitution, says that "so far, social stature has been based on labels that cover only part of our individual profiles, such as academic background, place of work, place of birth and income. It may be possible to create a fairer, more efficient society than what we now have by replacing these labels with evaluations of all aspects of a person."
On the other hand, he warned that "these evaluations may lock in discrimination that lasts a lifetime. Moreover, we risk creating a dangerous society that violates individuals' inner selves and controls them, prompting everyone to move in the same direction."
In the age of the internet, technology has functioned more to expand personal freedoms by enabling people to communicate anonymously or to use social media to present themselves in multiple ways. In Shishido's analysis, however, technology in the age of AI is likely to integrate one's different "faces" and create a measurable representation of individuals' entire personas.
A society based on a new sort of class system will arise if this measurement attains a certain level of importance.
"There is not enough debate on how this danger can be avoided. Before such systems are introduced lock, stock and barrel, we need to first think about the kind of society we are seeking to create," Shishido said.
The quantification of an individual's creditworthiness and the use of such information is nothing new. However, both the data on which the calculations are based and the areas in which they were applied were limited. Lawyer Takafumi Ochiai, who heads a subcommittee at the Fintech Association of Japan, explains that FICO -- the information used to determine credit scores in the United States -- is limited to a prospective client's history of payments for credit cards and loans, along with outstanding balances.
The scoring system used in China has attracted attention because the data used to calculate scores and the application of these scores extend to every corner of life. Zhima Credit, a subsidiary of the Alibaba group, uses as information not only one's affiliations and assets, but also evaluates potential customers' history of online shopping on Alibaba platforms and cashless purchases, social media connections and data owned by affiliate companies. Scores are used not only for credit, but also for employment, marriage, educational advancement, visa applications, hotel reservations and bicycle rentals, among other applications.
Yusaku Nishimura, a professor at the University of International Business and Economics in China, feels that "Chinese people use this [system] well." Various benefits are available for those with high scores. On the other hand, those with low scores will have difficulty even finding a job.
One matter of concern is how the Chinese government will utilize this system with its social credit system. Aiming to achieve the centralization of personal credit information by 2020, some cities are already registering those guilty of illegal parking or not paying public transport fares.
A system of rewards and penalties is being introduced. Those with low scores will be forbidden from using public transport, while those who have contributed to society will pay cheaper fares. According to Nishimura, some of the model cities for this system, such as Hangzhou in Zhejiang Province and Suqian in Jiangsu Province, have started to use scores compiled by private companies such as Zhima, connecting to their platforms.
-- Information bank
An information bank collects various data from individuals and sells it to companies that want to use it. Benefits for individuals can be assumed to potentially include data management and reward points. In order to secure the credibility of such services, the Information Technology Federation of Japan launched an accreditation system and started accepting applications at the end of last year.
Read more from The Japan News at https://japannews.yomiuri.co.jp/
