Showing posts with label Data science. Show all posts
Showing posts with label Data science. Show all posts

Saturday, September 23, 2023

Review: Maladies of Empire by Jim Downs

Maladies of Empire tells the usually-untold story of how epidemiological and medical advances were made directly as a result of the institutional aspects of colonialism. The “told” story usually includes Onesimus, an enslaved man who told his master, an 18th century New England priest by the name of Cotton Mather, about the practice of inoculation against smallpox using exposure to a small quantity of the virus to prevent a more severe, systemic infection. Mather tested this process on some 250 enslaved people and eventually his own son, convincing the aggressively skeptical medical establishment of the validity of this Black “folk” wisdom. Good discussion of this episode in history is usually limited to the ethicality of experimentation on enslaved people (see Medical Apartheid). But there’s sort of a comforting aspect to this narrative too: Onesimus was a relatively well-treated slave, and that his master listened to him and learned from his cultural traditions plays into the story the Anglo empire likes to tell that cultural mixing was beneficial for itself and all the civilizations it conquered.

Where Maladies of Empire goes beyond this is to document how the very processes through which the British Empire colonized much of the world also enabled the medical field to understand the spread of diseases. The story starts with slave ships:market forces drove slavers to keep costs down as much as possible without hurting the sale price of their human wares. This captive population was carefully documented and experimented upon, and from here, the medical establishment learned about the minimum fruit or vegetable intake required to stave off scurvy.

The mechanisms of the spread of infectious disease, however, necessitated the aggregation of records from across the world. Who was living where? Where did these ship passengers come from? Where did they go? After how many days did symptoms start? The meticulous records through which the British Empire tracked their ships and military and subjects allowed medical doctors to track the spread of yellow fever and cholera. For the first time, scientists were analyzing data that others had collected, possibly from the other side of the globe. 

The birth of epidemiology is therefore also, in a sense, the birth of data science, and the dark sides of data science were present from the start. The individual names and stories of the people who contributed the data — often racialized or institutionalized or poor — are lost to the sands of time, while the knowledge gleaned from their data goes on to benefit the wealthy and white. Downs’ story-telling is up for a challenge: how do you bear witness to these lost narratives and humanize the individual subjects whose suffering taught us how to cure or prevent disease, without getting mired in details? I don’t think the result is fully successful — there were some episodes where I felt the main themes became a little lost in the weeds of names and locations. But the work is excellent for understanding how intimately linked the development of science was with imperialism.

We see similar beats today: the bureaucracy of institutionalized people supports medical advances. For example, the link between the Epstein-Barr virus and Multiple Sclerosis was shown quite definitively only because of the mandatory monitoring and testing of American military recruits (themselves an imperializing force).

Downs contrasts the racism of the British Empire in the 19th century with that of the United States. In general, the British were certainly white supremacist, but more accepting of belief systems that allowed for similarities between races. For example, see Florence Nightingale and some of her peers’ views of racial differences in disease susceptibility:

Although Florence Nightingale believed in racial difference, regarding the English as the finest race on the planet, she did not use race as an explanation for the spread of cholera or other infectious diseases. Even after germ theory became widely accepted, she insisted that unsanitary environments led to disease. She did not believe that the source of disease transmission could be found in innate characteristics of the patient (...). Similarly, while Gavin Milroy and other doctors working in the Caribbean certainly harboured racist beliefs, they too searched for the cause of disease in the natural and built environment. Milroy condemned Black people’s living conditions and blamed their high rate of illness on their failure to maintain clean homes, but he did not focus on racial difference as the cause of disease spread.

Because their economic system depended on enslavement (and later, subjugation and segregation) of the Black race, American doctors approached medicine quite differently, and sought to reify the impact of race in health. The answer to “why is disease more prevalent in slaves?” could not be that they were oppressed, and forced into terrible living conditions, since that was a threat to the social order:

Many doctors in other parts of the world were turning to the physical world and the built environment to understand how disease spread; they observed symptoms in a patient and then turned outward to housing, sewers, drainage, and crowded conditions to understand why patients were sick. USSC surgeons did the opposite. They turned inward to the patient, trying to find the answer to the illness within or on their body. While they considered the natural or built environment, they emphasized racial identity as the cause. 

This approach had a long-lasting impact on the medical establishment: while slavery ended with the Civil War, “the USSC resurrected slave-holding ideologies to amplify racial difference and to contribute to medical knowledge.” These were not the first scientists to seek to justify their pre-existing beliefs with “evidence” and refuse to consider alternative explanations, and they were certainly not the last.

A challenge with books of this sort is where they stop. The British Empire is no more, but the world is still scarred by imperialism. Science has developed into a far more robust practice, but is still often racist, and the fruits of its research are unequally distributed. The author set out to tackle this topic for a reason, and I would imagine it is because he saw similarities between this part of history and our world today. If so, I agree, and I have highlighted some of these themes above. But Downs never goes so far as to explicitly draw out the link, to comment on practices of the twentieth century and beyond. I suppose it is the careful conservative nature of most academics, who don’t dare step outside their field of expertise — but that just leaves me, with my considerably smaller extent of expertise, to apply what I’ve learned on my own.

Friday, May 20, 2022

Review: "Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor" by Virginia Eubanks

Something about the digital world frightens people into believing they are facing a completely alien, Cthulian beast, as opposed to simply an online version of the usual suspects. Shoshana Zuboff makes this mistake in The Age of Surveillance Capitalism, suggesting the 'behavioural surplus' extracted by Facebook and its ilk fuels an economic system fundamentally different from the wholesome, warm and fuzzy capitalism of Henry Ford and company. I wrote in my review of that book that she failed to substantiate her argument:

Is Google hiding how much data it collects from you really all that different from Apple hiding the conditions of its manufacturing facilities? Is Facebook's attempts to manipulate your emotions or your sense of self-worth really a whole new beast or just another step in the advertising industry's development? Is the desire of surveillance capitalism companies to expand vertically and horizontally into new parts of our lives and into new parts of the world, to privatize or profit off public goods any different from the same expansion drive of any other company?

Virginia Eubank's Automating Inequality sees through the Silicon Valley smoke and mirrors, and instead correctly draws a direct line from the poorhouses of the 19th century, through the scientific charity and eugenics of the 20th century to the automated and algorithmic social systems of today. She coins the term "digital poorhouse", likening the publicly-funded facilities that granted wretched living conditions in exchange for grueling work to the systems of digital tracking and automated decision-making that govern distribution of public resources today.

Like the brick-and-mortar poorhouse, the digital poorhouse diverts the poor from public resources. Like scientific charity, it investigates, classifies, and criminalizes. Like the tools birthed during the backlash against welfare rights, it uses integrated databases to target, track, and punish.

She tracks three systems in particular: IBM's "modernization" of the welfare administration system in Indiana, the social sorting algorithm implemented for sheltering the unhoused in Los Angeles, and a model implemented in Pittsburgh to predict child harm. The chapters detailing these examples are compelling, and combine stories from social workers and people affected by these systems with data and perspectives from academics. They're also infuriating and saddening to read.

The final chapter, in which she ties together these stories with the cultural practices that enable them to exist (e.g. culture of individuality, middle class anxiety, racism) is excellent. Eubanks founds her critique of these systems in historical understanding of how these systems came to be.

Just as the county poorhouse was suited to the Industrial Revolution, and scientific charity was uniquely appropriate for the Progressive Era, the digital poorhouse is adapted to the particular circumstances of our time. The county poorhouse responded to middle-class fears about growing industrial unemployment: it kept discarded workers out of sight but nearby, in case their labor was needed. Scientific charity responded to native elites' fear of immigrants, African Americans, and poor whites by creating a hierarchy of worth that controlled access to both resources and social inclusion. Today, the digital poorhouse responds to what Barbara Ehrenreich has described as a "fear of failing" in the professional middle class.

I think because she is able to see the similarities between current technological solutions and social systems of the past, she is better able to identify the unique aspects of modern automation and algorithms. She concludes that the digital poorhouse is hard to understand, massively scalable, persistent over time, and is alienating in a particularly new way:

Containment in the physical institution of a poorhouse had the unintentional result of creating class solidarity across race, gender, and national origin. When we sit at a common table, we might see similarities in our experiences, even if we are forced to eat gruel. Surveillance and digital social sorting drive us apart as smaller and smaller microgroups are targeted for different kinds of aggression and control. When we inhabit an invisible poorhouse, we become more and more isolated, cut off from those around us, even if they share our suffering.

Working in data science, I think often about the ethical obligations of the profession. Sometimes I wish books like this one (along with Cathy O'Neil's Weapons of Math Destruction and Caroline Criado Perez's Invisible Women) were required reading. I'm under no illusion that professional certification or licensing of data science would solve the issue. Eubanks isn't, I think, the first to suggest a Hippocratic Oath for data science. Perhaps that would help with a culture shift.

I'll end with her two questions she asks people developing technological solutions that address poverty, because I think they're great:

  1. Does the tool increase the self-determination and agency of the poor?
  2. Would the tool be tolerated if it was targeted at non-poor people?

Thursday, December 16, 2021

Review: The Age of Surveillance Capitalism by Shoshana Zuboff

Briefly, Zuboff's accounts of the horrors of surveillance capitalism (clandestine collection of data and manipulation of behaviors) and the landmark legal cases surrounding this industry are well-documented but won't contain new stories for anyone already familiar with the topic.

This book isn't without a few good points. The repeated mantra of "who knows? who decides? who decides who decides?" is a good starting point for media/PR criticism. I also agree with Zuboff that using only terms like "monopoly" and "privacy" as grounds for criticizing surveillance capitalism industry leaves us woefully unprepared for battling surveillance capitalism and the commodification of human behavior.

I was hoping she would deliver on the promises she made in the intro: to demonstrate that surveillance capitalism was a separate beast from regular old capitalism. And while she uses terms that make it seem like the nuts and bolts of surveillance capitalism are distinct from capitalism (e.g., "behavioral surplus", "prediction imperative"), I think she really fails at making this argument. Is Google hiding how much data it collects from you really all that different from Apple hiding the conditions of its manufacturing facilities? Is Facebook's attempts to manipulate your emotions or your sense of self-worth really a whole new beast or just another step in the advertising industry's development? Is the desire of surveillance capitalism companies to expand vertically and horizontally into new parts of our lives and into new parts of the world, to privatize or profit off public goods any different from the same expansion drive of any other company? If anything, Zuboff inadvertently convinced me the exact opposite is true: surveillance capitalism is just capitalism.

Sunday, October 11, 2020

Review: Invisible Women by Caroline Criado Perez

 Rating: 4/5 stars

Invisible Women: Data Bias in a World Designed for Men  is a great overview of systemic biases that harm women. It's a little like a cultural version of the more biological Delusions of Gender; it's a well-organized, well-referenced, approachable/conversational synthesis of a broad range of sexist beliefs/sexist structures. Criado Perez did a fantastic job in tracing how small biases or assumptions, power imbalances in who makes decisions, individual choices, etc, translate into much broader social injustices. I also appreciated that she would highlight positive efforts and the benefits they've already brought; it's easy to criticize something, and harder to fix it.

This book was not without it's own biases. European languages were the emphasis of studies like how language shapes perception of gender. Western countries, or charities/NGOs based in western countries, made up nearly all the examples of pro-women initiatives; many of the examples of injustices in western countries were very much "white collar" sexism, while examples of sexism in developing countries were very much centered around sanitation issues. Despite being the home for one seventh of the world's women, China was virtually ignored. Despite housing another seventh of the world's women, discussion of women's issues in India was largely limited to the availability of public toilets - certainly a very, very, very crucial issue... but in contrast, the life of a woman in the UK was sliced and diced in pretty much every way imaginable.

The chapter on politics was, I thought, a little disappointing. The author advocates for women taking up a broader proportion of parliamentary bodies, citing, as far as I can tell, as single study that found that governments with more women pass legislation that promotes education and healthcare. Her summary of this study doesn't describe whether these women are from progressive/left wing parties, but later the author indeed describes how in many countries, the conservative/right wing parties have poorer representation of women. One might imagine that it is not the balance of women, but the overall political persuasion of the government that is the causal factor in both a pro-education/pro-healthcare policy as well as a more gender-balanced governing body. The author translates this into a shakily-founded critique of Bernie Sanders/ advocacy for Hillary Clinton (c. 2016 democratic primary). According to the author, Clinton would, by virtue of her gender, be a better leader for women - with no investigation of the two leaders' policies, nor even a glance at the gender make-up of their aides, assistants and advisors (the people who typically write the legislation proposed by the leaders and interpret the legislation voted on by the leaders). I'm a little bit exhausted of reading liberal feminists working through their 2016 Democratic Primaries grief, but at least this section wasn't as bad as Down Girl: The Logic of Misogyny. Still, this type of bias makes me question the author's interpretation of some of the fields she discusses that I am less familiar with - and that's a shame!

Despite these flaws, I think it is a book worth recommending and worth reading, and even those familiar with literature on how medicine, technology, law, etc propagate injustice against women will probably come away with something new.

I read the audiobook, which was narrated by the author, and her reading was enjoyable.