Eubanks - the digital poorhouse

Author & Context
- Virginia Eubanks, author of "The Digital Poorhouse,"
- Associate Professor of Political Science at University at Albany, SUNY.
- Founding member of Our Data Bodies project.
- Discusses transformation in decision-making from humans to machines over the last 40 years.

Historical Context
- Four decades ago, major life decisions (employment, mortgages, etc.) were made by humans.
- Humans used actuarial processes, but discretion prevailed.
Current Landscape
- Decision-making largely automated, with systems controlling important life aspects:
  - Policing, resources allocation, employment short-listing, fraud investigations.
- Obvious tools: Cameras, GPS.
- Invisible tools: Algorithms, social media data, embedded in life activities.
Transparency Issues
- Limited public access to algorithms shaping life chances.
- No sunshine laws for digital decision-making disclosures (except credit reporting).

Marginalized Groups Targeted
- Data scrutiny is not uniform; targets social groups (e.g., people of color, migrants).
- The most marginalized experience heightened data collection:
  - Public benefits access, policing, healthcare systems, border crossing.
Reinforcement of Marginality
- Data targeting increases scrutiny and perpetuates existing inequalities.
- Feedback loop of injustice affects marginalized populations disproportionately.

2014 TANF Benefits Incident
- Targeting families receiving cash benefits under TANF program.
- Data mined from EBT card transactions (ATMs in smoke shops/liquor stores).
- Publicly disclosed transactions to imply TANF fraud, despite constituting only 0.3% of cash withdrawals.
Legislative Response & Consequences
- Proposed laws requiring TANF recipients to provide detailed cash receipts.
- Unconstitutional bills aimed to stigmatize welfare recipients.
- Similar targeting of low-income groups across the nation.

Impact on Public Assistance Programs
- Automated eligibility systems discourage legitimate claims for benefits.
- Predictive models label struggling parents as "risky".
- Insufficient privacy safeguards in data collection for the vulnerable.
Technological Optimism vs. Reality
- Technology proponents argue that automation creates efficiency.
- Nevertheless, programs supporting the poor face increasing unpopularity due to stigma.

Physical vs. Digital Poorhouse
- Past containment in institutions created solidarity among the poor.
- Today's digital systems create divisions and targeted aggression against specific groups.
- This reflects long-standing American traditions of managing poverty rather than eradicating it.

Historical Context
- Contrast between old physical systems (county poorhouses) and modern digital methods.
- Evolution from physical containment strategies to digital surveillance.
Scalability of Digital Poorhouse
- Digital systems can rapidly expand and involve millions of applications.
- Examples include public assistance applications in Indiana and housing systems in Los Angeles.

Concept of Rational Discrimination
- Explanation by Oscar Gandy; lacks bias but reinforces existing inequalities.
- Automation intensifies structural inequalities when biases are ignored.
Historical Comparison
- Removal of discretion (e.g., mandatory minimums in criminal justice) hasn’t alleviated racial disparities.

Middle-Class Vulnerability
- Potential for middle-class individuals to become entangled in the digital poorhouse.
- Need for an awareness of changing socioeconomic dynamics.
Culture of Scrutiny
- Digital systems extend scrutiny beyond the poor, affecting the middle class as they slip down socioeconomic ladders.

Societal Reflection
- Society has long constructed systems that marginalize vulnerable populations.
- Computing and surveillance technology could exacerbate inequalities unless actively countered.
Call for Change
- Emphasis on the need for explicit commitment to justice and equality.
- Necessity of purposefully building an alternative system focused on equity.