Knowledge Vault 2/63 - ICLR 2014-2023
Ruha Benjamin ICLR 2020 - Invited Speaker - 2020 Vision: Reimagining the Default Settings of Technology & Society
<Resume Image >

Concept Graph & Resume using Claude 3 Opus | Chat GPT4 | Gemini Adv | Llama 3:

graph LR classDef king fill:#f9d4d4, font-weight:bold, font-size:14px; classDef power fill:#d4f9d4, font-weight:bold, font-size:14px; classDef race fill:#d4d4f9, font-weight:bold, font-size:14px; classDef tech fill:#f9f9d4, font-weight:bold, font-size:14px; classDef justice fill:#f9d4f9, font-weight:bold, font-size:14px; A[Ruha Benjamin
ICLR 2020] --> B[King: overinvesting in tech,
underinvesting morally. 1] A --> C[Binary view of tech:savior or destroyer. 2] A --> D[Power shapes machine learning. 3] D --> E[Race and tech mutually shape. 4] A --> F[Imagination is contested battleground. 5] F --> G[Crime alerts shape safety perceptions. 6] F --> H[Fintech exploits borrowers via algorithms. 7] F --> I[Films illustrate techno-utopian duplicity. 8] F --> J['Racist robots' stories evolve. 9] F --> K['Race-neutral' algorithm reproduces disparities. 10] A --> L[New Jim Code: racism encoded as progress. 11] L --> M[Conceptual frames: inequity, discrimination,
exposure, benevolence. 12] L --> N[Recidivism surveys encode racism. 13] L --> O[Surveillance renders groups invisible
or hypervisible. 14] L --> P[Corporate diversity ensures
innovation contains. 15] L --> Q[AI hiring mirrors discrimination. 16] Q --> R[Job seekers feel dehumanized,
develop subversive tactics. 17] A --> S[Tech worker organizing challenges
harmful practices. 18] A --> T[Racial literacy initiatives in tech. 19] T --> U[Community orgs advance tech justice. 20] U --> V[Coalition halts predictive policing in St. Paul. 21] A --> W[Arts and humanities vital for examining tech. 22] W --> X['White Collar Crime Early Warning'
parody subverts policing. 23] W --> Y[Creative projects expose
embedded discrimination. 24] A --> Z[Each tech twist is chance
to weave new patterns. 25] Z --> AA[Vast injustice motivates
becoming 'pattern makers'. 26] Z --> AB[Historical approach can
encode justice. 27] Z --> AC[Critical traditions develop
justice strategies. 28] A --> AD[Support tech justice,
embed equity in innovation. 29] A --> AE[Just tech requires wrestling
power, context, imagination. 30] class B king; class C,D,E power; class F,G,H,I,J,K,L,M,N,O,P,Q race; class R,S,T,U,V,W,X,Y tech; class Z,AA,AB,AC,AD,AE justice;


1.-Dr. Martin Luther King warned about overinvesting in technology and underinvesting in social and moral development. We need to develop technology humanely and justly.

2.-We tend to see technology in binary terms - it will save or destroy us. We need to recognize human agency in shaping technology.

3.-Power dynamics, not just ethics, shape machine learning. Racism is productive, generating things of value to some while harming others.

4.-Race and technology mutually shape each other. Social norms and values precede and influence what technologies are deemed desirable and inevitable.

5.-Imagination is a contested battleground that influences technology and social order. Many live inside the nightmares enabling elite fantasies.

6.-Apps like Citizen broadcast crime alerts, shaping perceptions of public safety while ignoring issues like reckless driving. Social anxieties shape technologies adopted.

7.-Fintech lenders in Kenya entice and exploit borrowers, harvesting digital data to infer character. Algorithms scoring individuals significantly shape economic prospects.

8.-The film Sleep Dealer illustrates the duplicity of techno-utopian fantasies - easy access to some is a nightmare for others.

9.-News stories about "racist robots" have evolved from shock to unsurprise to attempts to override biases. Differentiating technologies that differentiate us is challenging.

10.-A study found a "race-neutral" healthcare algorithm reproducing racial disparities by using costs to predict needs. Indifference to social reality is harmful.

11.-The New Jim Code concept examines how racism gets encoded in technology under the guise of progress. Related concepts include coded bias, algorithms of oppression.

12.-Four conceptual frames: engineered inequity (explicitly seeks to amplify divisions), default discrimination (ignores cleavages, reproducing defaults), coded exposure (surveillance vs inclusion), techno-benevolence (addresses bias narrowly).

13.-Recidivism risk assessment surveys encode structural racism without mentioning race. Colorblind inputs still produce biased outputs.

14.-Surveillance technologies render racialized groups either invisible or hypervisible in ways that enable vulnerability. Creative resistance emerges.

15.-The show Better Off Ted satirizes how a superficial corporate diversity ethos and whiteness of tech development ensures innovation produces containment.

16.-AI hiring tools aim to reduce bias but can mirror discriminatory patterns against black and female applicants. Technical fixes appear desirable but perpetuate harms.

17.-Job seekers feel dehumanized by AI interviews with no transparency. Some develop subversive tactics like using invisible keywords to game automated screenings.

18.-Labor organizing among tech workers has challenged corporate collaboration with harmful state practices, connecting to a longer activist history.

19.-Educational initiatives develop racial literacy in tech, examining how structural racism operates in technologies and building capacity for intervention.

20.-Community organizations like Data for Black Lives and Detroit Community Tech Project advance proactive approaches to tech justice through policy and grassroots education.

21.-In St. Paul, a coalition successfully organized against a predictive policing program targeting youth, dissolving it for a community-led approach.

22.-The arts and humanities are vital for critically examining technologies. Imagining alternative realities helps see current injustices.

23.-A parody "White Collar Crime Early Warning System" subverts predictive policing by flagging likely sites of financial crimes by corporate executives.

24.-Creative projects upset the status quo to expose discrimination embedded in technologies by reversing common targets.

25.-If inequity is woven throughout society, each twist and code of technology is a chance to weave new patterns and politics.

26.-Accepting the vastness of injustice motivates becoming "pattern makers" who actively create new possibilities through technology.

27.-An ahistorical approach to machine learning can capture and harm; a historically and sociologically grounded approach can encode justice and empowerment.

28.-Critical intellectual traditions continually develop insights and strategies for justice that can be built upon in the context of new technologies.

29.-The speaker calls for supporting tech justice initiatives and finding ways to build on the tradition of embedding equity in innovation.

30.-Developing technology justly requires wrestling with power dynamics, social and historical context, and the contested terrain of imagination.

Knowledge Vault built byDavid Vivancos 2024