Skip to main content
MIT AI Risk Initiative
BackErasing social groups
Home/Risks/Shelby et al. (2023)/Erasing social groups
Demeaning social groups
Alienating social groups

Erasing social groups

Sociotechnical Harms of Algorithmic Systems: Scoping a Taxonomy for Harm Reduction

Shelby et al. (2023)

SourceDOI
Sub-category
Risk Domain
1Discrimination & Toxicity
1.3Unequal performance across groups

Accuracy and effectiveness of AI decisions and actions are dependent on group membership, where decisions in AI system design and biased training data lead to unequal outcomes, reduced benefits, increased effort, and alienation of users.

people, attributes, or artifacts associated with specific social groups are systematically absent or under-represented... Design choices [143] and training data [212] influence which people and experiences are legible to an algorithmic system(p. 728)

Entity— Who or what caused the harm

Human

Due to a decision or action made by humans

AI system

Due to a decision or action made by an AI system

Other

Due to some other reason or is ambiguous

Intent— Whether the harm was intentional or accidental

Intentional

Due to an expected outcome from pursuing a goal

Unintentional

Due to an unexpected outcome from pursuing a goal

Other

Without clearly specifying the intentionality

Timing— Whether the risk is pre- or post-deployment

Pre-deployment

Occurring before the AI is deployed

Post-deployment

Occurring after the AI model has been trained and deployed

Other

Without a clearly specified time of occurrence

Part of Representational Harms

Other risks from Shelby et al. (2023) (24)

Representational Harms

1.1 Unfair discrimination and misrepresentation
OtherUnintentionalPost-deployment

Representational Harms > Stereotyping social groups

1.1 Unfair discrimination and misrepresentation
AI systemUnintentionalPost-deployment

Representational Harms > Demeaning social groups

1.1 Unfair discrimination and misrepresentation
AI systemUnintentionalPost-deployment

Representational Harms > Alienating social groups

1.1 Unfair discrimination and misrepresentation
AI systemUnintentionalPost-deployment

Representational Harms > Denying people the opportunity to self-identify

1.1 Unfair discrimination and misrepresentation
AI systemUnintentionalPost-deployment

Representational Harms > Reifying essentialist categories

1.1 Unfair discrimination and misrepresentation
AI systemUnintentionalPost-deployment
View all 24 risks from this paper →