DATA C104 Midterm 2: Case Studies & Readings

Case study: Sewing machine

The sewing machine mechanized sewing, promising to free women from manual labor and meet clothing needs. Tech narratives overhyped liberation to drive buy-in, shaping perceptions of progress while ignoring class and gender inequalities.

HCE tools:
labor – shifted women’s work from home-based unpaid labor to industrial wage labor, intensifying exploitation rather than reducing it.
identity/positionality – middle-class women saw it as liberation, while working-class women faced harsher factory conditions; gender and class shaped who benefited.
narratives – “liberation through technology” story naturalized women’s productivity expectations and obscured structural inequality.

Reading: Diana’s Onlife World by Mireille Hildebrandt

Hildebrandt’s story shows a world where smart technologies like Diana’s PDA “Toma” seamlessly manage daily life — anticipating needs, making decisions, and regulating behavior through invisible data-driven systems. These systems blur the line between online and offline life, subtly shaping human action. Laws are public, contestable, and allow disobedience. Tech rules are invisible and automatic, often deciding for users without consent.

HCE tools:
agency – human agency is constrained as Toma makes preemptive choices; Diana’s control over decisions is diminished by automated systems.
power – technological systems govern behavior through design and data rather than law, exerting power that’s invisible and hard to contest.
sociotechnical system – Diana’s life illustrates how data, algorithms, and human routines are intertwined in one continuous, self-regulating system.

Reading: The Carrier Bag Theory of Fiction by Ursula K. LeGuin

Le Guin reframes storytelling and evolution around cooperation and sustenance rather than conquest. The “bag” — a tool for holding and sharing — replaces the “spear” as the symbol of culture, valuing collective creation over domination.

HCE tools:
narratives – challenges the dominant “Hero Story” that glorifies violence and progress, offering an alternative story centered on care and interdependence.
representation – reclaims the overlooked roles of gatherers and makers, reshaping who and what counts as culturally significant.
co-production – the shift in storytelling mirrors and shapes a shift in social values, showing how ideas of technology and society evolve together.

Reading: The Doomsday Invention by Raffi Khatchadourian

Khatchadourian explores philosopher Nick Bostrom’s warning that superintelligent AI could surpass human control, posing an existential threat. The “intelligence explosion” could create systems whose goals conflict with human survival, making preparation an urgent moral task.

HCE tools:
power – superintelligent AI concentrates immense decision-making power beyond human oversight, reshaping global hierarchies of control.
agency – human agency is threatened as AI systems gain autonomous, self-improving capabilities that act on their own logic.
narratives – the “doomsday” story frames AI as an inevitable, apocalyptic force, shaping public fear and policy around technological destiny.

Reading: Hypernudge by Karen Yeung

Yeung describes how Big Data enables “hypernudges” — constantly updated, personalized digital environments that steer user behavior without awareness. These algorithmic feedback loops erode autonomy and reshape social and political life.

HCE tools:
power – corporations wield covert power by shaping choices through data-driven design rather than explicit control or consent.
agency – human decision-making is subtly constrained as algorithms predict and guide behavior in real time.
sociotechnical system – user data, algorithms, and digital platforms interact continuously, forming a self-reinforcing system of behavioral regulation.

Case Study: Personality Tests and the Making of Self

From WWI military screening to modern MBTI apps, personality testing has shaped how people understand themselves. These systems classify individuals and feed back social meaning, turning selfhood into something produced through data and interaction.

HCE tools:
classification – organizes people into fixed personality types, defining identity through standardized categories.
identity/positionality – people internalize and perform identities shaped by test results and social feedback.
performativity – personality categories don’t just describe traits—they help create the very selves they claim to measure.

Reading: Making Up People by Ian Hacking

Hacking shows how scientific and social classifications don’t just describe people—they create new kinds of people. Through the looping effect, labels like “multiple personality” or “autism” shape behavior and self-understanding, which then reshape the classification itself.

HCE tools:
classification – diagnostic labels organize people into social and medical categories that redefine identity and experience.
performativity – classifications bring new kinds of people into being by making the label socially and psychologically real.
co-production – science and society evolve together as new knowledge about people changes how they live and are understood.

Case Study: “guns don’t kill people”

The debate over gun agency challenges the idea of technological neutrality. Latour argues that humans and artifacts together form hybrid actor-networks—once a person holds a gun, both are transformed, and new actions become possible.

HCE tools:
agency – the gun and human share agency; technology shapes and enables human action rather than merely serving it.
co-production – human intentions and technological capacities evolve together, creating new social realities.
power – distributing agency to artifacts reveals how technology can embody and enact political and moral power.

Case Study: door closers and speed bumps

Latour shows how mundane technologies embody and enforce social rules. Door closers “remember” to shut doors, and speed bumps “make” drivers slow down—delegating human intentions to objects that act in our absence.

HCE tools:
agency – objects exercise delegated agency by performing human tasks and shaping behavior.
co-production – social order and technology are made together as norms become built into material forms.
sociotechnical system – humans and objects form a single system where social expectations and technical functions continuously reinforce each other.

Reading: Where Are the Missing Masses? by Bruno Latour

Latour argues that artifacts are the “missing masses” of society—nonhuman actors that carry intentions, enforce norms, and shape how people behave. Technologies like door-closers or seat belts embed moral and social prescriptions into their design.

HCE tools:
agency – artifacts act as agents by delegating and enforcing human intentions through design.
co-production – social order and technological design evolve together, each shaping the other’s form and function.
sociotechnical system – humans and artifacts form intertwined networks where behavior, responsibility, and control are distributed.

Case Study: COINTELPRO

The FBI’s Counterintelligence Program used surveillance, infiltration, and disinformation to target civil rights and anti-war activists, exposing how state surveillance enforces control and suppresses dissent.

HCE tools:
power – surveillance technologies concentrated political power in the state, enabling covert control over marginalized groups.
identity/positionality – activists were targeted based on race and political affiliation, revealing how surveillance amplifies social inequality.
representation – state data practices constructed activists as “threats,” shaping public perception and justifying repression.

Case Study: Muslim Surveillance Post-9/11

After 9/11, U.S. surveillance expanded through the PATRIOT Act and NSA programs, disproportionately targeting Muslim Americans. Monitoring of mosques, phone data, and apps exposed how security technologies deepen social and political divides.

HCE tools:
identity/positionality – Muslim identities were constructed as inherently suspicious, making certain groups more exposed to state control.
power – surveillance systems extended state authority under the guise of security, concentrating control over marginalized populations.
vulnerability – targeted surveillance increased exposure to harm and discrimination, showing how technology distributes risk unequally.

Reading: The Most Surveilled Place in America by Gaby Del Valle

Del Valle shows how border surveillance and “Prevention Through Deterrence” policies push migrants into deadly desert routes. The U.S.–Mexico border has become a militarized surveillance zone where human suffering and environmental damage are built into the system.

HCE tools:
power – surveillance and border technologies extend state control and violence under the guise of security.
vulnerability – migrants bear the greatest physical and social risks, revealing how surveillance systems unevenly distribute harm.
sociotechnical system – drones, sensors, and walls form a network linking state policy, industry profit, and human movement into one deadly infrastructure.

Reading: Contextual Integrity by Helen Nissenbaum

Nissenbaum redefines privacy as maintaining the proper flow of information within its social context. Privacy violations occur when data is shared or used in ways that break contextual norms or undermine the purpose of that setting.

HCE tools:
classification – determines what counts as appropriate information within each context and who has access to it.
representation – data flows create representations of people that may misalign with social expectations or intended purposes.
power – those who control data flows (platforms, institutions) shape norms of privacy and can override contextual boundaries.

Case Study: Havasupai Blood Samples

The Havasupai tribe gave blood for diabetes research, but scientists used the samples for unrelated studies on ancestry and mental illness. The misuse exposed how data extraction can violate collective identity and Indigenous sovereignty.

HCE tools:
representation – genetic data stood in for tribal identity, producing narratives about ancestry without community consent.
power – researchers and institutions exercised colonial authority over Indigenous bodies and knowledge.
identity/positionality – the tribe’s experience shows how data practices intersect with history, culture, and power, shaping who controls self-definition.

Case Study: Crisis Text Line

Crisis Text Line used machine learning to support people in moments of distress, but later shared anonymized data with a for-profit partner. This repurposing of sensitive crisis data exposed ethical gaps around consent and profit in digital care.

HCE tools:
vulnerability – users shared data in moments of crisis, making them especially exposed to exploitation and harm.
contextual integrity – data collected for care was used in a new commercial context, violating the norms of the original setting.
power – corporate interests benefited from user data without accountability or fair distribution of value.

Case Study: TSA Semi-Automated Body Scanners

After 9/11, the TSA introduced body scanners that reveal detailed images for security screening. While official safeguards (blurred faces, image deletion) address data protection, they overlook deeper issues around bodily privacy and social norms.

HCE tools:
contextual integrity – airport screening changes what counts as “appropriate” exposure of the body, redefining privacy in a security context.
norms and embodiment – scanners are built around binary body assumptions, causing discomfort or discrimination for non-binary and trans travelers.
power – security technologies normalize invasive practices in the name of safety, shaping how bodies are seen and governed.

Reading: The Right to Oblivion by Lowry Pressly

Privacy isn’t just about controlling personal data — it’s about protecting oblivion, the right not to be known or recorded at all. Oblivion preserves freedom and the open-endedness of human life.

HCE tools:
oblivion vs secrecy – secrecy hides existing facts; oblivion stops facts from ever being created.
critique of informational privacy – when privacy is framed as data control, it still feeds surveillance and capitalism by assuming everything can be captured and managed.
historical lens – mass documentation (like photography) turned everyday life into permanent records, sparking modern privacy fears.
freedom – true privacy protects the ability to disappear, remain undefined, and resist total legibility.

Case Study: Logical Positivism and Its Limits

Logical positivists believed only statements that can be logically or empirically verified count as meaningful knowledge. But Hume’s Problem of Induction exposed a key flaw: no amount of evidence can ever guarantee a universal truth — only that it hasn’t yet been disproven. The discovery of a single black swan was enough to overturn centuries of “verified” observation.

HCE tools:
verification vs falsification – knowledge can’t be proven true forever, only tested until disproven.
empiricism’s limits – observation is powerful but incomplete; science depends on uncertainty, not certainty.
knowledge and humility – the “black swan” reminds us that claims to total truth often ignore what lies outside experience.

Case Study: Falsification and Its Challenges

Popper argued that science progresses by testing and potentially disproving theories. However, real-world science faces challenges: old theories persist despite contrary evidence, experiments depend on multiple assumptions, and different theories can explain the same observations.

HCE tools:
falsifiability – theories must be testable and disprovable to count as scientific.
theory-ladenness of experiments – failed tests don’t always indicate which part of a theory is wrong (Duhem-Quine problem).
underdetermination – multiple theories can explain the same data, complicating the idea of definitive scientific truth.

Reading: Introduction A Role for History by Thomas S. Kuhn

Kuhn argues that science doesn’t progress linearly but through revolutions. Normal science reinforces existing paradigms, while anomalies eventually trigger a shift to a new, incompatible framework, transforming the scientific worldview.

HCE tools:
paradigm and normal science – scientific research operates within accepted frameworks that shape what questions are asked and which answers are considered valid.
scientific revolutions – knowledge evolves discontinuously when anomalies overthrow prevailing theories.
historical perspective – past scientific beliefs were coherent in their own context, showing that what counts as “truth” depends on historical and social factors.

Reading: Reimagining Open Science Through a Feminist Lens by Denisse Alejandra

Alejandra argues that open science often reproduces power imbalances. A feminist approach emphasizes care, collaboration, and context-sensitive practices to make openness equitable and responsible.

HCE tools:
power – global North institutions dominate open science, shaping who benefits and who is marginalized.
identity/positionality – recognizing diverse actors, including Indigenous and local communities, ensures knowledge practices are inclusive.
co-production – science, ethics, and social context evolve together; equitable open science requires integrating care, accountability, and local knowledge into research design.

Case Study: AIDS Crisis

During the AIDS epidemic, slow government action and rigid biomedical rules led the gay community to organize, advocate for faster drug approvals, and conduct independent trials, demonstrating the value of lay expertise.

HCE tools:
agency – the community exercised independent action and influence on research and policy outside formal institutions.
co-production – activism and biomedical science shaped each other, showing how societal needs drive scientific practices.
power – traditional regulatory systems held authority over drug approval, but grassroots efforts redistributed influence and challenged institutional control.

Case Study: Anthropogenic Climate Change

Since the 1970s, scientific consensus has shown humans drive global warming. Yet media framing and industry-funded misinformation delayed public understanding and policy action.

HCE tools:
power – fossil fuel interests and media outlets shaped public perception, amplifying doubt despite scientific consensus.
representation – scientific findings were misrepresented to appear uncertain, influencing public belief and policy.
narratives – competing stories (consensus vs. doubt) framed climate change as controversial, affecting societal response.

Case Study: Italian Lab Technician Earthquake Prediction

Giampaolo Giuliani used radon gas measurements to predict an earthquake, but seismologists dismissed him due to lack of credentials. When a major quake struck, standard scientific procedures were criticized for failing to warn the public.

HCE tools:
agency – Giuliani and the seismologists illustrate how scientific authority mediates who can act and who is heard.
power – institutional and expert authority shaped whose predictions were taken seriously and whose were ignored.
co-production – scientific norms, expertise, and public safety expectations interact, showing how knowledge, responsibility, and societal outcomes are mutually shaped.

Reading: The Science of Lambs by Harry Collins and Trevor Pinch

After Chernobyl fallout contaminated sheep in Cumbria, scientists and farmers clashed over risk assessment and response. Farmers’ local knowledge was ignored, and official explanations were inconsistent, undermining trust and livelihoods.

HCE tools:
agency – farmers exercised practical expertise, but scientific authority often constrained their influence on decisions.
power – government and scientific institutions controlled messaging, regulations, and risk definitions, shaping outcomes for local communities.
co-production – scientific models, local knowledge, and policy interacted, showing how technical assessments and social context jointly produce the management of environmental crises.

Case Study: Layers in the Standard SV History

Silicon Valley developed through successive layers: early electronics and military tech, Stanford’s academic-industry connections, the semiconductor boom, and successive computing waves, producing the modern tech ecosystem.

HCE tools:
sociotechnical system – innovation emerged from intertwined networks of universities, industry, and technology.
co-production – technological development and social/economic structures evolved together, shaping the Valley’s culture and economy.
power – military, academic, and entrepreneurial institutions directed which technologies thrived and who benefited from them.

Reading: Eisenhower Speech

Eisenhower warned that the military–industrial complex could concentrate power, threatening democracy and public oversight. He emphasized that citizens must remain vigilant to ensure military strength serves peace, not entrenched interests.

HCE tools:
power – the military–industrial alliance concentrates political, economic, and scientific influence.
agency – citizen awareness and action are necessary to check the influence of powerful institutions.
co-production – science, government, and industry shape one another, creating systems that can either support liberty or entrench elite control.

Reading: Hail the Maintainers by L. Vinsel and A. Russel

Vinsel and Russel argue that society glorifies “innovation” while undervaluing maintenance work. Repair, upkeep, and caregiving sustain technology and society, often performed by women and marginalized groups, and are essential to real progress.

HCE tools:
labor – maintenance work is socially organized, essential, yet undervalued, reflecting inequalities in who does and benefits from it.
power – cultural focus on innovation privileges entrepreneurs and tech elites while obscuring the contributions of maintainers.
narratives – the “innovation myth” shapes perceptions of progress, masking the critical role of sustaining existing systems.

Reading: The Silicon Valley Loop by Malcom Harris

Harris shows that Silicon Valley follows a repeating cycle of hype, crash, and reinvention, concentrating wealth and power among the same founders and investors. Platform companies prioritize growth and monopoly, often offloading costs onto workers.

HCE tools:
power – capital and influence are concentrated in the hands of repeat investors and tech elites.
labor – workers and users bear the social and economic costs of platform growth strategies.
sociotechnical system – financial, technological, and social structures interact to reproduce cycles of speculation and inequality.

Case Study: TCA Contamination in San Jose

Toxic solvent from Fairchild Semiconductor contaminated neighborhood drinking water, causing health risks and birth defects. Lorraine Ross organized the community, leading to environmental advocacy and Superfund designation.

HCE tools:
power – corporate and industrial practices created environmental hazards, while regulatory authority determined accountability.
vulnerability – local residents, especially children, were exposed to harm, revealing unequal risk distribution.
agency – community organizing and activism transformed residents from affected parties into actors shaping policy and remediation.

Case Study: Amazon Echo

The production and use of Amazon Echo devices create widespread social and environmental impacts. Mining for minerals exploits labor in the Global South, generates pollution, and produces e-waste, while cloud computing adds significant energy and water demands.

HCE tools:
labor – extraction and production rely on exploitative labor practices, disproportionately affecting marginalized populations.
vulnerability – communities in the Global South bear environmental, health, and social risks of tech production.
sociotechnical system – supply chains, data centers, and consumer use form an interconnected system distributing environmental and social costs unevenly.

Reading: Anatomy of an AI System by Kate Crawford and Vladan Joler

Crawford & Joler show that AI devices like Alexa depend on global networks of material extraction, labor, and data. What seems “immaterial” in the cloud is built on human exploitation, environmental harm, and resource-intensive supply chains.

HCE tools:
labor – workers in mining and manufacturing supply chains face exploitative conditions that enable AI production.
vulnerability – people and ecosystems at each stage of the supply chain are exposed to environmental and social harms.
sociotechnical system – AI’s functionality relies on intertwined networks of material, labor, and data infrastructure, showing intelligence is distributed across humans, machines, and the environment.

Case Study: Data as a Commodity

Framing data as “the new oil” positions it as a natural resource to be extracted and traded, obscuring that data is socially produced and context-dependent. Its value is created through surveillance, prediction, and targeted advertising.

HCE tools:
representation – data stands in for human behavior and preferences, shaping markets and social understandings.
power – corporations control collection, interpretation, and monetization, concentrating influence over individuals and society.
performativity – treating data as a commodity shapes both business practices and the behaviors it measures, creating a feedback loop that enforces its economic value.

Reading: Digital Capitalism by Jathan Sadowski

Sadowski argues that digital capitalism turns everyday life into data to be collected, analyzed, and monetized. Companies link online and offline behaviors to predict and control people, concentrating economic and social power.

HCE tools:
power – control over data flows gives corporations economic dominance and influence over social behavior.
representation – personal actions are converted into data points that stand in for people, enabling prediction and control.
vulnerability – constant data extraction exposes individuals to surveillance, manipulation, and loss of autonomy.

Reading: OpenAI and Exploited Labor by Billy Perrigo

Perrigo highlights that ChatGPT’s “safe” AI depends on low-paid Kenyan workers labeling toxic content, often under traumatic conditions. The human labor behind AI is hidden from users, raising ethical concerns beyond algorithmic design.

HCE tools:
labor – low-wage workers perform essential, emotionally taxing tasks that enable AI functionality.
vulnerability – workers face physical, emotional, and economic risks, disproportionately impacting marginalized populations.
sociotechnical system – AI depends on the interplay of algorithms, human labor, and global labor infrastructures, showing intelligence is distributed across humans and machines.

Case Study: First Industrial Revolution

Work shifted from homes to factories, where artisanal skills were mapped into machines. Time and labor were disciplined and measured, setting the stage for mechanization and automation.

HCE tools:
labor – work became structured, hierarchical, and measurable, distinguishing skilled from unskilled roles.
power – factory owners and managers controlled labor through timekeeping, machinery, and production processes.
sociotechnical system – machines, human labor, and organizational practices were intertwined, producing industrial efficiency and control.

Reading: Labor by Kate Crawford

Crawford shows that AI depends on human labor at every stage, from warehouse workers to low-paid data labelers. Automation reshapes work but doesn’t eliminate it, often making human effort invisible, timed, and devalued.

HCE tools:
labor – human work underpins AI, often undervalued, monitored, and exploitative.
vulnerability – workers face economic, physical, and emotional risks, especially in low-paid or invisible roles.
sociotechnical system – AI’s functionality arises from intertwined human labor, machines, and organizational practices, masking the labor behind apparent automation.

Case Study: Uber

Uber classifies drivers as independent contractors, shifting costs and risks onto them while avoiding labor protections. Platform technology extracts maximum efficiency from workers’ time and vehicles, creating flexible but precarious labor.

HCE tools:
labor – drivers perform essential work under precarious conditions without benefits or protections.
power – Uber controls work conditions and labor classification, concentrating economic and legal authority.
sociotechnical system – platform algorithms, vehicles, and worker behavior interact to maximize efficiency while externalizing costs.

Reading: It’s Not Technology That’s Disrupting Our Jobs by Louis Hyman

Hyman argues that job instability results from corporate and policy decisions, not technology itself. Industrialization and modern gig work reflect social and economic choices that shape labor conditions.

HCE tools:
labor – work arrangements, job security, and exploitation are socially structured, not determined by technology.
power – corporations and policymakers shape labor conditions, controlling who benefits and who is precarious.
co-production – technology and social/economic structures evolve together; industrial and digital tools are embedded within broader labor systems.