Short answer

Actively seek out and integrate diverse cultural knowledge systems into the design and development of AI to mitigate inherent biases and create more representative technologies.

Field: Classic Design
Source: Deep Blue (University of Michigan) (2022)
Method: Ethnocomputing, Critical Discourse Analysis, Archival Research, Art-Based Research
Evidence: Strong effect

Current AI development is heavily influenced by Western, colonial perspectives on knowledge, potentially marginalizing or erasing non-Western ways of knowing. This classic design research insight is drawn from a 2022 study published in Deep Blue (University of Michigan). Using Ethnocomputing, critical discourse analysis, archival research, art-based research, researchers explored how this design variable affects real-world outcomes. The key design takeaway: Actively seek out and integrate diverse cultural knowledge systems into the design and development of AI to mitigate inherent biases and create more representative technologies.

Study

Classic DesignHigh ImpactStrong effect

Algorithmic Bias Reflects Historical Colonial Epistemologies

Current AI development is heavily influenced by Western, colonial perspectives on knowledge, potentially marginalizing or erasing non-Western ways of knowing.

Deep Blue (University of Michigan) · 2022

Key Findings

01AI's notion of 'intelligence' is often based on a singular, European colonial definition of knowing.
02Data production processes carry deep epistemological biases that favor colonial perspectives and efface Black worldviews.
03This bias creates an 'epistemic gap' that limits technical approaches and scholarly expressions in AI.
04An anticolonial practice is needed to redress data for African and African diasporic cultural knowledge in AI.

Application

Design takeaway

Actively seek out and integrate diverse cultural knowledge systems into the design and development of AI to mitigate inherent biases and create more representative technologies.

How to apply

When designing AI systems, consider the origin and potential biases of the datasets used. Explore methods for co-designing with communities whose knowledge is traditionally underrepresented in technology.

Project actions

01When researching a design problem, consider the historical and cultural context of the existing solutions or technologies.
02Investigate the origins of the data or information you are using for your design project to identify potential biases.

Method & Evidence

AimHow do historical colonial epistemologies embedded in data production shape contemporary Artificial Intelligence, and what are alternative approaches to AI development that incorporate marginalized cultural knowledge systems?

MethodEthnocomputing, Critical Discourse Analysis, Archival Research, Art-Based Research

ProcedureThe research analyzes historical archives and contemporary AI projects, comparing Western epistemologies with those of African and African diasporic heritage, specifically Gullah Geechee culture, to identify biases and propose alternative computational frameworks.

ContextArtificial Intelligence Development, Cultural Heritage, Data Science

Variables

IVHistorical colonial epistemologies embedded in data production

DVBias and limitations in contemporary AI systems, marginalization of non-Western knowledge

CV["Specific cultural knowledge systems (e.g., Gullah Geechee)","Contemporary AI development practices"]

Strengths & Limitations

Strengths

+Interdisciplinary approach combining ethnography, computer science, and cultural studies.
+Focus on underrepresented cultural perspectives in AI.

Limitations

It can be challenging to access and interpret historical data from non-dominant cultures accurately.

Reliability & validity

The validity of the findings relies on the thoroughness of archival research and the depth of ethnographic engagement. Reliability can be enhanced through triangulation of data from multiple sources and methodologies.

Think critically

How can designers actively decolonize their design processes and the technologies they create, moving beyond simply identifying bias to actively embedding diverse epistemologies?

Design Principles

"Design for epistemological diversity: Ensure that design processes and outputs acknowledge and incorporate a multiplicity of knowledge systems, moving beyond dominant cultural paradigms."

Understanding the historical and cultural roots of data and algorithmic design is crucial for creating more inclusive and equitable AI systems. Designers must critically examine the underlying assumptions and values embedded in the tools and data they use.

What This Means for Your Design

The way computers 'learn' (AI) is often based on old ideas from European colonialism, which might not understand or value how other cultures know things. This can lead to unfair AI. We need to include different cultural knowledge to make AI better for everyone.

How to use in your project

1.Reference this research when discussing the societal impact of technology or the importance of diverse data in your design project.

Add to My Project

Quick Cite

Paragraph starter

This research highlights how Artificial Intelligence often reflects dominant, colonial epistemologies, creating an 'epistemic gap' that marginalizes non-Western ways of knowing. Designers must critically examine the cultural biases embedded in data and algorithms to develop more equitable and inclusive technological solutions.

Source

Deep Blue (University of Michigan)

Ancestors and Algorithms: Ethnocomputing AI with African & African Diasporic Heritage

journal · 2022

View source

Questions About This Research

What does the research say about algorithmic bias reflects historical colonial epistemologies?: Actively seek out and integrate diverse cultural knowledge systems into the design and development of AI to mitigate inherent biases and create more representative technologies. Evidence: Deep Blue (University of Michigan) (2022).
Why does "Algorithmic Bias Reflects Historical Colonial Epistemologies" matter for design?: Understanding the historical and cultural roots of data and algorithmic design is crucial for creating more inclusive and equitable AI systems. Designers must critically examine the underlying assumptions and values embedded in the tools and data they use.
How can designers apply this research?: Actively seek out and integrate diverse cultural knowledge systems into the design and development of AI to mitigate inherent biases and create more representative technologies.
What were the main findings?: AI's notion of 'intelligence' is often based on a singular, European colonial definition of knowing.. Data production processes carry deep epistemological biases that favor colonial perspectives and efface Black worldviews.. This bias creates an 'epistemic gap' that limits technical approaches and scholarly expressions in AI.. An anticolonial practice is needed to redress data for African and African diasporic cultural knowledge in AI.
What research method was used?: Ethnocomputing, Critical Discourse Analysis, Archival Research, Art-Based Research.
How strong is the evidence?: Evidence strength is rated Strong effect, based on a 2022 journal from Deep Blue (University of Michigan).
What should I do differently in my next project?: When designing AI systems, consider the origin and potential biases of the datasets used. Explore methods for co-designing with communities whose knowledge is traditionally underrepresented in technology.
What are the limitations?: The study's focus on specific African diasporic cultures may not fully represent the diversity of all marginalized knowledge systems.