Social Credit: Identity and Authority

Jesse Hirsh

easyDNS is pleased to sponsor Jesse Hirsh‘s “Future Fibre / Future Tools” segments of his new email list, Metaviews

Who gets to say who you are?

 
It’s been over a month since the last issue in our social credit series, and yet it’s a topic I think about often. In particular I’m still of the belief that social credit systems will be a lasting response to this crisis, especially given increasing concern about identity.

On the one hand there will be the issue of one’s relationship with the novel coronavirus. Do you have it? Have you had it? Are you infectious? Are you immune?

However on the other hand, over the past couple of weeks we’ve seen a new metric emerge: are you a protester breaking curfew? Are you a threat to security (or private property)? Is your speech dangerous? Are you sharing dangerous information online?

We still tend to regard the world through the lens of the 20th century, assuming we live in an analog world when everything is digital now. Data is not only easy to collect, in many cases it is mandatory or automatic.

Perhaps we realize that everything we do online is tracked, but what about the actions we take offline? Or maybe we’re never actually offline given the super computers many of us carry in our pockets? Similarly our interactions with institutions are almost always recorded, creating greater depth and reach to our data trails.

This kind of information quickly becomes unmanageable without some form of synthesis or summation. Scores and rankings become necessary if any of this information is to be used or understood.

Social credit systems are also essential for algorithmic applications. Think of them as the cause of and solution to algorithms and AI managing the world. Since AI cannot think for itself, it needs an external basis for assigning value and determining priorities, and that’s where scores and ranking play an essential role.

Technology scholars, policy-makers, and executives in Europe and the United States disagree violently about what the digitally connected world should look like. They agree on what it shouldn’t: the Orwellian panopticon of China’s Social Credit System (SCS). SCS is a government-led initiative to promote data-driven compliance with law and social values, using databases, analytics, blacklists, and software applications. In the West, it is widely viewed as a diabolical effort to crush any spark of resistance to the dictates of the Chinese Communist Party (CCP) and its corporate emissaries. This picture is, if not wholly incorrect, decidedly incomplete.

This is a crucial point, and one of the reasons why we’re producing our social credit series. Much more work needs to be done to understand how these systems work. We also need to keep testing the hypothesis that they’re also active outside of China, especially in North America.

SCS is the world’s most advanced prototype of a regime of algorithmic regulation. It is a sophisticated and comprehensive effort not only to expand algorithmic control, but also to restrain it. Understanding China’s system is crucial for resolving the great challenges we face in the emerging era of relentless data aggregation, ubiquitous analytics, and algorithmic control.

Regulatory systems often operate as a paradox. On the one hand they seek to restrain and control, but on the other hand they tend to legitimize and expand that which they seek to regulate.

Social credit systems are a great example of this, as they’re intend to place limits on systems by institutionalizing ranking and reputation, however those ranking and reputation systems legitimize and expand the system in question. Obvious examples are Uber and Airbnb which both used ranking and reputation to legitimize and expand people using their personal property to compete in commercial marketplaces.

Although this comes with a whole new set of problems, including the exercise of arbitrary authority:

A major concern about SCS and similar systems is that they will not actually operate on the basis of consistent rules. One of the problems with algorithmic regulation generally is that algorithmic decisions may not be amenable to explanation and interpretability in the same way as human ones. Administrative law is based on the principle of reasoned explanation for decisions to restrain agency overreach or arbitrariness. That may be difficult to square with systems, such as those employing AI techniques of deep learning, which make decisions based on hidden patterns of correlations with no direct analogue to explanations humans can understand.

A recurring theme within our initial research into social credit systems is that transparency is essential. Gaming the system seems to be an inherent risk to any social credit application, and transparency has to be part of mitigating that.

Although on a meta level, transparency may be crucial to enabling us to understand how and what the social credit system is in general:

China’s Social Credit System resembles the proverbial elephant approached by a series of blind people. Each individual touches a separate location, and believes he or she is encountering a different animal. SCS is China’s version of the Western credit reporting system, built in the internet era rather than the mainframe era. It is also the Chinese government’s attempt to replace inefficient and sometimes corrupt human enforcement of the law with consistent machine-based systems, promoting public management and societal trust. At the same time, it could be a dangerous fusion of the worst aspects of surveillance capitalism and Chinese digital authoritarianism. It draws on Western principles of scientific management as well as Confucian ideals of social harmony. These are only a few of the ways SCS might appropriately be viewed.

I find this analogy of the elephant being explored by blind people (or in the dark) as applying to the digital monopolies in general. We know that they’re massive, but we fundamentally do not understand the extent of their influence and power.

The same can be said about our current pandemic. We know that it’s massive, but do not (as yet) understand it’s full impact and reach. Once again the problem is a lack of transparency:


Although one could also argue, that even with some transparency, our collective resolve to fight this pandemic may be failing:

Americans may wish the virus to be gone, but it is not. While the outbreak has eased in the Northeast, driving down the overall national numbers, cases have only plateaued in the rest of the country, and they appear to be on the rise in recent days in COVID Tracking Project data. Twenty-two states reported 400 or more new cases Friday, and 14 other states and Puerto Rico reported cases in the triple digits. Several states—including Arizona, North Carolina, and California—are now seeing their highest numbers of known cases.

These numbers all reflect infections that likely began before this week of protest. An even larger spike now seems likely. Put another way: If the country doesn’t see a substantial increase in new COVID-19 cases after this week, it should prompt a rethinking of what epidemiologists believe about how the virus spreads.

But as the pandemic persists, more and more states are pulling back on the measures they’d instituted to slow the virus. The Trump administration’s Coronavirus Task Force is winding down its activities. Its testing czar is returning to his day job at the Department of Health and Human Services. As the long, hot summer of 2020 begins, the facts suggest that the U.S. is not going to beat the coronavirus. Collectively, we slowly seem to be giving up. It is a bitter and unmistakably American cruelty that the people who might suffer most are also fighting for justice in a way that almost certainly increases their risk of being infected.

There are a number of different scenarios arising from these recent developments.

I do think there is hope or the possibility that epidemiologists will revise their models and that the virus does not spread as they had feared. It remains dangerous, and difficult to contain, but may not require the kind of total lock down we initially faced.

The concern however is whether the general population can handle such nuance, or whether they will go from one extreme to the other. From total lock down to having no fucks to give.

The danger there is that the virus could once again spread quickly. And let us not forget this virus is absolutely brutal for many who catch it:

This is why we have to remain concerned and vigilant. This is not over hyped. You do not want to catch this thing.

Allow me to take a moment to share with you my current speculation as to how this virus operates:

While the virus targets the respiratory tract, the disease it enables appears to be neurological. While it can take hold in the lungs, it can impact almost any aspect of your body, depending on your immune system and vulnerabilities.

It does not seem to be transmitted outdoors, but does flourish in contained areas with poor air circulation. The higher your exposure, i.e. the more virus molecules you consume, the more severe your potential reaction. Limited exposure can result in limited effects and high exposure can result in severe effects.

While asymptomatic carriers are less likely to transmit, pre-symptomatic carriers can have high amounts of viral shedding (i.e. producing and sharing virus molecules via their breathing and spit). Obviously anyone with symptoms, whether coughing or sneezing, may be shedding high volumes of virus molecules as well.

The infectiousness of a single infected individual may not be as high as previously believed, but a group of people in an enclosed space can provide conditions for the virus to spread quickly and with severity.

Again, this is just my reading of available research, and as a layperson, an attempt to translate what more informed people are saying and arguing.

My point however in presenting this take, is to point out how nuanced and relatively complicated this pandemic picture has become. We’re not dealing with an easy binary of infected or not infected, or vulnerable vs immune. Instead we’re facing a situation where multiple vectors of needs and concerns intersect.

Which brings us back to the issue of who you are and what status you have? Automated contact tracing was one proposed means of helping to decipher that status, and thankfully it is being resisted by health authorities who recognize that manual tracing, while more labour intensive, is a far more effective solution.

Although one of the benefits of contact tracing is that it helps map out the social connections, whether formal or informal, strong or weak, that fuel the spread of this virus.

Which is why I keep coming back to social credit systems, as a potential response to this pandemic, and the growing data that it is both producing and potentially requiring.

Do you have it? Have you had it? Do you know or have had contact with people who have it? Did you shop at a store with an outbreak? Did you protest with people that were infected? Are you infectious? Are you immune? What’s your pandemic score?

Who gets to say who you are? Who measures your relationship to the pandemic? Or the revolt?

Related Posts

Defund Big Tech, Refund Community

I got an email last week from a long time fellow traveler, Andrew Clement, pointing me to a new paper that he had participated in producing.

Defund the police?!

I’ve spent the better part of the last decade arguing that institutional authority has been replaced by cognitive authority. While not mutually exclusive, the argument focuses on how people build trust, and why they choose to trust who they do (or why they don’t). Specifically that institutions are no longer the basis of trust, but how people communicate (and what they do) is far more important.

GPT-3 and existential angst

In yesterday’s issue we looked at GPT-3 from a technical perspective, in today’s issue let’s delve into the philosophy surrounding it. After all, technology as we know it today is as much philosophical or even ideological as it is technological.

Leave a Reply

Your email address will not be published. Required fields are marked *