Linked from

The 42 pages that link to Algorithmic bias, each with the reason it gives.

• Artificial intelligenceRelated: AI can reproduce disparities present in its data, objectives, or deployment.

• StereotypeRelated: Models can reproduce stereotypes reflected in training data or design choices.

• Social mediaRelated: Ranking and recommendation systems can distribute visibility unevenly across users and groups.

• Content moderationRelated: Automated moderation can disproportionately flag or overlook particular groups’ expression.

• Machine learningRelated: Training data and modeling choices can reproduce or amplify unequal treatment.

• DiscriminationRelated: Automated decisions can reproduce discrimination through data, design, or deployment.

• Digital humanitiesRelated: Biased data and models can distort historical representation and computational interpretation.

• RecidivismRelated: Risk tools may reproduce disparities when trained on unequal patterns of arrest or punishment.

• Disparate impactRelated: Automated decisions may create measurable disparities even when their rules omit protected traits.

• Financial inclusionRelated: Automated credit decisions can reproduce exclusion through biased data or models.

• Forensic psychologyRelated: Automated risk assessments may reproduce disparities in the data and institutions that shape them.

• Platform governanceRelated: Automated enforcement can apply rules unevenly across groups and languages.

• Wearable technologyRelated: Sensor performance and training data can produce less reliable results for some groups.

• Recommender systemRelated: Training data and ranking choices can distribute recommendations unevenly across groups.

• Business intelligenceRelated: BI can reproduce biased records or measures when organizations treat analysis as neutral.

• Institutional discriminationRelated: Automated decisions can reproduce institutional disparities through data and design choices.

• Smart cityRelated: Automated city decisions can reproduce unequal treatment across communities.

• Age verificationRelated: Automated age estimates can perform unevenly across demographic groups.

• Large language modelRelated: Patterns and stereotypes in training data can shape unequal model responses.

• Online datingRelated: Ranking and recommendation systems may reproduce unequal visibility or treatment among users.

• AI ethicsRelated: Bias raises questions about fairness in data, model design, and automated decisions.

• Quantitative dataRelated: Biased measurements or samples can carry inequities into numerical models and decisions.

• Social networking serviceRelated: Ranking and recommendation can distribute visibility unevenly across users and communities.

• Big dataRelated: Patterns in biased or unrepresentative data can be reproduced at scale.

• Google MapsRelated: Uneven map coverage or ranking can shape which places gain visibility.

• Insurance fraudRelated: Uneven fraud flags can burden groups whose data resemble patterns labeled suspicious.

• No-free-lunch theoremRelated: The theorem clarifies that algorithmic preferences are necessary sources of conditional advantage.

• Rotten TomatoesRelated: Selection and classification rules shape which reviews count and how reception is represented.

• Information systemRelated: Automated system decisions can reproduce or amplify biases in data and design.

• Fourth Industrial RevolutionRelated: Data-driven decisions can reproduce discrimination when models or deployment practices encode unequal treatment.

• Health informaticsRelated: Health algorithms can reproduce inequities when their data or design reflects existing disparities.

• SmartwatchRelated: Sensor models may perform unevenly across skin tones, bodies, and activities.

• Technology and societyRelated: It shows how social inequalities can persist inside automated decisions.

• ChatbotRelated: Training data and response policies can produce unequal chatbot treatment across groups.

• CyberfeminismRelated: Biased systems reveal how gender inequality can persist inside apparently neutral computation.

• Philosophy of artificial intelligenceRelated: It tests whether AI decisions can be fair when data and design encode disparities.

• Biographical evaluationRelated: Automated analysis of biographies can reproduce inequalities embedded in historical data.

• Dead Internet theoryRelated: Algorithmic influence can distort visibility even when the content itself is human-made.

• Emerging technologiesRelated: Bias can become embedded in emerging systems before their design and effects are well understood.

• Generative AIRelated: Training data and model design can reproduce or amplify social inequalities.