First off, I would like to extend my heartfelt gratitude for the great comments and questions I’ve been getting from classmates. These have given me much to think about, as did points made in our cohort meeting in week 3.

Some highlights from last week that I wanted to highlight because I found something especially sparking in them that’s stuck with me after some days lapsed:

Janel on measuring unemployment: “I’ve just been thinking about how measuring ‘jobs’ and ‘unemployment’ is constantly changing and difficult to make reasonable conclusions over time.” <- This is a valid point, and I don’t think necessarily limited to how we measure jobs and unemployment but how the nature of jobs and unemployment are transforming over time, whether it’s through technologies that make coordinating large amounts of gig work possible or our definitions (like recognition of “underemployment,” where a worker has a labor-wage job that doesn’t give them enough hours to meet all their monetary needs/wants, as a legitimate problem that “unemployment” fails to account for).

Darcie on my idea to study international migrants in southwestern China: “…can you explore a different country? Not sure if the same problems exist in other global regions but I think migration is happening all over the place.” <- This has stuck with me because I wonder if it’s possible. Part of what fascinates me about this group isn’t so much the group itself (which likely exists in other countries that border Myanmar), but the act of migration coming up against a system depending increasingly on all-consuming digital surveillance and which also “others” the migrants (for example, this week, I–a migrant in the same system but likely inhabiting a more privileged position–was temporarily locked out of both of my private Wechat and Alipay wallets, one because the system decided my paying my gas bill which I pay every month to a state-backed utility company was suspicious and the other simply because my Chinese green card which I use for ID places me in a “foreign” box in which all members, it seems, are assumed to be potential scammers in need of extra policing). What’s going on is that migrants are separated from citizens inside the national ID system that is used in identity verification. The Chinese government maintains a centralized database of Chinese IDs, but it’s basically outsourced verification of migrants/foreigners to private companies to deal with as they see fit. Furthermore, there are laws discouraging social media platforms from allowing accounts to exist anonymously, but the Chinese ID system has an API, so it’s easy to run verification checks against it. Usually, foreigners/migrants have to be verified by humans (as in a visual passport check) and a bank account name check (which obviously requires that one gets a bank account–not an easy thing to do in China unless you are employed by a company and have paperwork to back this up). Furthermore, there are so few foreigners in China (around 800K according to the 2020 Chinese census vs. 1.4B Chinese) that it doesn’t always make commercial sense for companies to go through the trouble of creating an in-house verification process for this group of people that is relatively expensive to verify. Thus you get a lot of platforms that don’t bother. This can have relatively minor effects (like a food order app that won’t give foreigners/migrants the same discounts that it will give Chinese people because its system won’t allow them to sign up to sign up for an internal eWallet, instead forcing them to use a 3rd party eWallet for which it doesn’t offer discounts) or it can have major possible problems like with this Yunnan health code that won’t allow foreign migrants under the age of 18 to be registered in the system which in turn affects whether or not they can gain access to things like hospitals or public transportation (many cities have policies in place for handling those without an electronic health code, but because these policies are actually executed on a hyper-local level by people who don’t have to have any training in such policies, foreign migrants are often just refused entry or service rather than being allowed to follow the rules that would apply to a Chinese citizen without a health code). It’s a headache in the urbanized, foreign-trade-heavy East where the foreign population is generally associated with foreign investment giving it some power to propel complaints through municipal governments. If you have a foreign migrant population without the “foreign investment” carrot, I don’t know how you get people to, say, allow you to enter a hospital without a health code. Most countries are not so reliant on eWallets and digitized health codes, but as things are pushed increasingly online, people without “correct” identification are in danger of getting marginalized (imagine if, say, a private hospital chain in the US decided it would use facial recognition to identify its customers, for example, so anyone not properly IDed would not be able to enter the hospital–it probably sounds ridiculous, but you do have buildings that bar the public to use toilets in large cities in the US, so I could see other forms of access restrictions implemented in a privatization push).

And Kate had good thoughts on trying to whittle topics down.

I listened to some things

In week 4, I went to several lectures. Highlights included one, which I think was backed by AI Ethics & Society at EFI was a talk by Dr. Milagros Miceli. She studies the impact these new AI technologies have on data workers and the roles data workers play to prop up these AI platforms. This was fascinating and terrifying, and she made a point that I’ve been mulling over for a while, that the powers running the platforms are perhaps understudied. Also that there is a disconnect between these people (think the people building the platforms/tools, the people with the jobs in need of dataworkers to power them) and this dataworking underclass that is trying to function within a very fixed framework (dataworkers don’t have the ability to dispute decisions made by the algorithms that accept/reject their work for example). It was absolutely fascinating and if you ever get the chance to see her speak, I highly recommend it. This paper she wrote was highly recommended by several audience members, but I haven’t had the chance to go through it yet. And then she also recommended this book by Mary Gray and Siddarth Suri: Ghost Work: How to Stop Silicon Valley from Building a New Global Underclass, which I’ve only just cracked open.

Then I mentioned The Algorithmic Imprint in my last blog. I had the chance to log on to a Q&A with one of the authors hosted by some grad students at Oxford, and he fleshed out his thoughts some more and also some background about how the study came about. Basically, the concept is when an algorithm (in this case, it was a grading algorithm for the British GCSE A-level exams because, during the pandemic, the exams were canceled, so the testing companies were trying to figure out a “fair” way of judging students without a big test) was implemented, it forced a change in the data collection system such that when people complained and the algorithm was removed, the system still ended up classifying people based on the data collected for the algorithm (the paper discusses this in-depth, but basically, Bangladesh’s students had a history of cramming for these tests in the last few months because the tests counted in ways that classes didn’t, so the students’ previous work in classes didn’t necessarily reflect how they would perform on the actual A-levels–the data collected for the algorithm didn’t take any of this into account which led to some wild ways of creating the data the testing companies were looking for). So it’s not enough to just remove an unfair algorithm if you don’t also deal with other parts of the system you have adjusted to accommodate the algorithm. Another thing that came out in the talk that I don’t think was included in the paper was that in the years following the rollout and revocation of the algorithm, even though the data collection never reverted back, people seemed to have accommodated and the low scores during the algorithm drama year improved back to levels similar to those before the algorithm was introduced. As Malcolm in Jurassic Park says, “Life finds a way.” Only in this case, replace life with Bangladeshi students. (for those that remember my cheerleading of the toilet article, this would count as another example in my book of people ultimately finding to adapt a system that was built in such a way that it overlooks them).

The final lecture that stuck out to me was one on racial justice and global public health by Dr. Ugo Edu hosted by the University of Washington Anthropology department. She struck on some things that I’d run across in the Built Infrastructure readings for the Exclusion and Inequality Intensive. In particular, a lot of her examples were about disconnects between organizations with money power and organizations they were trying to “help.” Things like a University in some country in Europe (I want to say it was one of the Scandinavian countries, but I could be misremembering) wanting to introduce a birthing technique that they said they had found to improve outcomes in a rural health clinic in some country in Africa. The staff at the health clinic were outraged when they discovered what the technique was: a variation on squatting while giving birth. The director of the health clinic refused the partnership, pointing out that that was how women had been giving birth locally until “experts” from the global north had come in and insisted that giving birth hitched up in stirrups in a hospital bed was the better way. There was another example where a country (didn’t mention where but also sounded like a global north country) wanted to provide aid money for contraceptives to a country in Africa. But when the country replied, “That’s great. We’ll be able to buy so many contraceptives from China with that money,” the country stipulated that the money could only be used to be contraceptives manufactured in the country providing the money (which of course would be considerably more expensive than the Chinese contraceptives and thus would help keep far fewer women from having unwanted pregnancies).

An article that came up that was interesting : a doctor in Switzerland wants to automate the process of deciding if someone after medically-assisted suicide is mentally fit enough to choose to die under a Swiss law that requires a psychological evaluation. The article goes into hospitals using algorithms to make other kinds of life and death decisions. If you read it, also read these two articales in AP and the Guardian about assisted suicide in Canada, and then tell me if this isn’t a creepy dystopia in the making. (Also, is just letting everyone kill themselves—and even potentially charging them to do it—if they’re too miserably impaired to function at what society deems “good” standards the most Capitalist answer to poverty ever?)

A topic update

As you’ve perhaps noticed, I keep finding myself gravitating toward how the makeup of who has the money/power in these systems heavily influences what does and doesn’t become implemented and who gets screwed over as a result. And so, I find myself gravitating toward my fourth topic suggestion from week 4: Is it possible to predict what climate policies will get passed by a legislative body by looking at the makeup of that body.

Some thoughts toward that end:

Project form: Part A) A simulation that takes in certain information about a policy and then runs a vote to decide if that policy will be passed in a certain legislature. If it fails, it states why it fails, given the information it has on what climate policies get passed within that legislature. Part B) Tracks different legislatures/congresses and policies coming up for a vote for any major changes that might change the voting outcome. Flags these and reports them. (this could ultimately be a tool to help reporters track climate policies in different countries) Part C) Paper on all the problems with the things the program(s) use(s) to make decisions—this is based on the idea that we examine less powerful groups in similar ways, for example, why might create predictions on refugees based on certain simplified variables even though every single refugee is a complex human being with a story with many different parts.

Possible questions to explore: Who is making policy decisions about climate change, how are they making those decisions, and can we use the history of legislatures/congresses to predict how likely dramatic climate change policy is to get passed in different governments around the world?

The idea is to turn how we study people and groups around to examine extremely powerful groups whose decisions hold the potential to marginalize and disempower a lot of people. I think I’m open to exploring other possible topics that might be better examined within this framework.

Toward that end, I discussed the viability of my building the simulator in Part A with a developer familiar with my strengths and limitations, and she thought it was feasible so long as I figure out a way to score the policies. There are considerably greater challenges with Part B (so this may or may not be doable in the time period depending on how much time it takes to solve all the challenges with Part A).

Power’s Discontents – Week 4/5 / Knowledge Integration and Project Planning: Data, Inequality and Society (2022-2023) by blogadmin is licensed under a Creative Commons Attribution CC BY 3.0