For as long as loans and debt have existed, the people who make loans care a great deal about how likely they are to get their money back. For a long time, the ways people acquired enough confidence to make a loan were informal — asking around, seeking references, only loaning to families with “a good name,” and so on. Then, in 1956, a financial company called Fair, Isaac, and Company introduced a systematic scoring system to rate the creditworthiness of U.S. citizens. The score was based on data about potential borrowers, rather than hunches about them. Over time, this process became the FICO score and pioneered the credit rating field, the business of systematically trying to work out how likely a person or a business is to pay back a loan.
The sources of data used to compile credit ratings were originally limited in nature and mainly related to internal records about prospective borrowers held by banks. During the information technology boom, these data grew in volume and detail, but remained fundamentally internal. They evolved into more and more complete list of what banks, credit card companies, and other similar entities knew about a person. However, as the internet revolution came to pass, some credit rating companies have begun to look at data created by people as they live their lives online. An average internet user’s “data trail” will contain public and private data on where they go, what they do, what opinions they express about various issues, and who they talk to. This information is potentially valuable to a financial company trying to decide if a person is likely to pay back a loan diligently or repeatedly miss their obligations.
Until recently, these kinds of data trails weren’t of much interest to governments, with the exception of security agencies that use data to catch certain kinds of criminals, such as gang members, terrorists, and pedophiles. In recent years, governments are starting to realize they might use this enormous wealth of personal data for purposes beyond the prevention and detection of classic criminal activities.
The most famous example of this is China’s “social credit” scoring system, which was announced in 2014 and expected to be rolled out by 2020 (Creemers 2018).
This system (or perhaps more accurately network of systems) develops profile scores for citizens, which are based on a range of activities that people carry out online and indeed offline. This score is then used to determine whether citizens are eligible to carry out a wide range of activities, such as buying domestic flight tickets or business class train tickets, or if they have full-speed access to the internet.
The social credit system is not entirely about deterring behaviors deemed to be undesirable. The Chinese government is also using good scores to give perks to citizens. These include avoiding the need for cash deposits when booking hotels and speeding up paperwork related to international travel. It is even claimed that dating sites are increasing prospective matches for people with higher social credit scores. Most notable for a study on citizen participation is the policy decision to give people with high credit scores preferential access to hospital doctors. This explicitly links social credit scores with the provision of public goods.
Most countries are far behind China in terms of explicitly analyzing online activity and generating a score from it. However, in a much softer way, this trend is already happening with Twitter’s Verified status to show someone is who they claim to be on Twitter and Facebook’s new Constituent Badges tool, a small service that shows a politician if a person leaving a comment or sending a message is actually a constituent.
Both these initiatives, which do not come with any of the reputational baggage of China’s credit scoring system, have the effect of telling decision makers that one person is more credible and more worthy of response than someone without that status. Neither system requires government intervention, nor has statutory status, but each still fulfills some of the same roles. These two interventions might be far removed from the Chinese social credit scoring system, but it is important to see the similarities between them.
Even in countries where state-sanctioned social credit scores will never become politically acceptable, an ever-increasing amount of data will likely be used by politicians and decision makers to help them identify which citizens are most worthy of their time.
Lest this seem like a dystopian vision driven solely by digital technology, it can be seen as the latest manifestation of a tradition of thought about citizen participation that goes back over a hundred years. This is a tradition that argues that enfranchisement, and the right to be listened to by government can be earned or lost through the behaviors of individuals. In the 19th century, the British philosopher John Stuart Mill advocated giving extra ballots to those who were more educated.6 In some countries, being incarcerated or having a criminal record is justification for withdrawing the right to vote.7This turns out to be highly problematic, of course, as differential rates of incarceration lead to differential enfranchisement of entire groups within countries, as found in the United States.
Not all choices to favor some groups over others are founded on a political philosophy. In most countries, politicians and other decisions makers will tend to give more attention to very wealthy citizens than those of more modest means. This is clearly driven by self- interest and homophily more than by ideology. Similarly, businesses use reputation data to segment and favor certain groups. Digital reputations on platforms like Amazon, eBay, and Wikipedia are used to give some people more power than others within those systems. Facebook recently admitted to rating users on a trustworthiness metric to help them automate anti-abuse measures (Dwoskin 2018).
What is not clear is the extent to which some countries will take deliberate actions to encourage, permit, or prohibit decision makers from taking into account social credit scores, or more informal online reputation scores. On the one hand, many countries have antidiscrimination laws that prevent governments from making decisions based on a citizen’s race, sexuality, or religious beliefs. On the other hand, it can be important for governments to have mechanisms to exclude citizens from services, positions, opportunities or physical areas. Motivation can vary from deeply legitimate concerns, such as preventing sex-offenders from working with children, to profoundly illegitimate ones, such as using records of private conversations to block people from employment.
The extent to which social scoring will be used to include or exclude people from influencing decision making is unknown. Some commentators are not optimistic. “One could argue that the dominant trend seems to be not so much a hierarchy in terms of who is listened to (the glass half full) but rather the actual trend looks more like the use of scoring to exclude (the glass more than half empty,” said American University Professor Jonathan Fox during an interview.
A lively debate will be seen soon in almost all countries about whether or not data emitted by citizens going about their everyday lives can and should be used to create scores that can include or exclude people from certain services or opportunities. Those countries that permit large amounts of data to be used to score and differentiate citizens will almost certainly see changes to the nature of power relationships that are highly likely to reinforce and exacerbate existing access and power inequalities.
6: Mill’s elitist proposal of weighing votes according to voter competence has gained renewed support. Two recent books revisiting the idea have received widespread attention in the mainstream media. In Against Democracy, the political scientist Jason Brennan (2016) proposes that voting be restricted to those who can pass a basic test on political knowledge. In her most recent book, Edge of Chaos: Why Democracy Is Failing to Deliver Economic Growth, the economist Dambisa Moyo (2018) also argues for a similar system in which the weight assigned to an individual’s vote could be determined by a civics test or one’s profession or educational level. ↩
7: For a comparative table across countries, see https://felonvoting.procon.org/view.resource. php?resourceID=000289. ↩