5AAVC300 Chap.6 Algorithmic Bias and Justice
Algorithmic Bias and Justice
Defining the problem
The Week 7 page defines algorithmic bias as systematic and repeatable errors in computer systems that create unfair outcomes, such as privileging one group of users over others.
It also asks how Black thinkers and activists combat algorithmic injustice as part of a longer abolitionist tradition.
The New Jim Code
Ruha Benjamin names technologies that reproduce existing inequities while being promoted as more objective or progressive the New Jim Code.
She starts from names and hiring, then moves to algorithms, databases, targeting and recognition systems, arguing that racism is an input to design as well as an output of failures.
Proxies and tailoring
Removing race from a system does not remove racial effects when location, viewing history or other variables act as proxies.
Tailored marketing can also become targeting and containment, and automated decisions can rebrand profit maximisation as bias minimisation.
From bias to justice
Benjamin treats private design choices as public policy and argues that private refusal is not a tenable answer for most people.
Her alternative is collective defiance, which links the week's two halves.
Essay angles for Week 7
Good questions ask whether bias can be fixed technically, whether race-neutral systems can be fair, and what algorithmic justice requires beyond better data.
Work from Benjamin's mechanisms, such as proxies and tailoring, and use the week page for the abolitionist frame rather than attributing it to her introduction.
What this chapter covers
- 01
Algorithmic bias as systematic, repeatable error
- 02
The New Jim Code and its name
- 03
Racism as an input to design
- 04
Proxies such as geography and viewing history
- 05
Tailoring that becomes targeting
- 06
Collective refusal and algorithmic justice
Worked example · free
Find the proxy in a hiring filter
- 2List the inputs and identify which could track race.
- 2Apply Benjamin's argument about proxies and coded inequity.
- 2Judge the neutrality claim and state what evidence would test it.
Key terms
- Coded Inequity
- Coded Inequity is Benjamin's term for inequality built into technical systems in ways that are faster and harder to challenge.
- Discriminatory Design
- Discriminatory Design is design that amplifies hierarchy, ignores and so replicates division, or tries to fix bias but worsens it.
- Proxy Variable
- A Proxy Variable is an input that stands in for a category a system claims not to use, such as location for race.
- Race Critical Code Studies
- Race Critical Code Studies is Benjamin's framework joining science and technology studies with critical race studies.
- Collective Defiance
- Collective Defiance is shared refusal of coded inequity, which Benjamin prefers to private opting out.
Algorithmic Bias and Justice FAQ
Is algorithmic bias just a data problem?
Partly, but Benjamin argues it is more. If racism enters design as an input, through purposes, categories and histories, then cleaner data cannot remove it alone. The fix has to reach who builds systems and for what ends.
Why does Benjamin call it the New Jim Code?
The name echoes Michelle Alexander's The New Jim Crow and the older Jim Crow, which began as a minstrel character in 1832. It signals continuity between earlier legal discrimination and newer technical systems presented as neutral or progressive.
Does hiring a more diverse team fix biased technology?
Benjamin calls a more diverse team an inadequate solution when discriminatory design grows from the interplay of racism and capitalism. Her Siri example shows a team member raising African American English and being told the products were for the premium market.
Assessment move
Make a clinic sheet with three rows, one for each kind of discriminatory design, and add an example from Benjamin to each. Practise the proxy test on two systems you know, listing every input and asking what it might stand in for. Then write a thesis on whether bias can be fixed technically, and end it with what collective action would add. Cite the Week 7 page whenever you use the abolitionist framing.
The week's other required reading is Cathy O'Neil's Civilian Casualties; note where her account of predictive models agrees or disagrees with Benjamin's.
Working through Algorithmic Bias and Justice in 5AAVC300? Sia is AskSia’s AI Arts and Humanities tutor — ask any 5AAVC300 Algorithmic Bias and Justice question and get a clear, step-by-step explanation grounded in how 5AAVC300 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.