서브메뉴
검색
Reimagining Privacy and Transparency in Visual Assistance Technologies with Blind People
Reimagining Privacy and Transparency in Visual Assistance Technologies with Blind People
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202105239
- ISBN
- 9798291568903
- DDC
- 004
- 서명/저자
- Reimagining Privacy and Transparency in Visual Assistance Technologies with Blind People
- 발행사항
- [Sl] : University of Michigan, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 146 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 87-03, Section: B.
- 주기사항
- Advisor: Brewer, Robin;Schoenebeck, Sarita.
- 학위논문주기
- Thesis (Ph.D.)--University of Michigan, 2025.
- 초록/해제
- 요약Artificial intelligence (AI) technologies are often advertised as a "solution" to remove accessibility barriers for blind people. At the same time, AI technologies are often designed without consulting blind people, exposing them to privacy harms and errors that are difficult to detect. Through 46 interviews and 16 focus groups with blind communities, this dissertation showcases that frameworks from disability justice and disability studies can guide the development of responsible AI systems. Specifically, this dissertation examines three interconnected projects around visual assistance technologies (VAT), which are real-world applications that blind people use to gain visual access. The first study explores the tensions of using AI to manage privacy concerns in VAT. Findings revealed that blind participants rejected automating privacy decisions, preferring greater control over when and how privacy is managed. Drawing from disability justice principles, this work generates design recommendations to build AI-enabled privacy tools that foster mutual understanding and collaboration. The second study explores how blind people detect and contest errors in VAT. Findings demonstrated that VAT often produces cultural biases and fails to support blind people in making sense of AI. By employing theories from disability studies, this research explains the contours of how VAT often reflects a limited representation of blind people's needs. The third study bridges the first and second projects, investigating how to support blind people in detecting errors in AI tools for privacy management. Findings indicated that transparency in AI-enabled privacy-enhancing technologies should encompass interaction-level cues and broader technical architecture, informing users of specific outputs and the overall system. The primary contributions of this dissertation are: (1) empirical accounts that detail the perspectives of blind communities on existing and emerging AI systems and affirm the central role blind people play in contesting, imagining, and verifying AI technologies, (2) theoretical implications that advance accessibility scholarship by applying disability studies and disability justice activism to reveal how blind communities reshape AI systems, and (3) design implications to expand the boundaries of responsible AI practice and co-create AI systems that honors the expertise of blind communities.
- 일반주제명
- Information technology
- 일반주제명
- Information science
- 일반주제명
- Computer science
- 일반주제명
- Disability studies
- 키워드
- Low vision
- 키워드
- Privacy
- 키워드
- Transparency
- 키워드
- Accessibility
- 키워드
- Blind
- 기타저자
- University of Michigan Information
- 기본자료저록
- Dissertations Abstracts International. 87-03B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017359945
■00520260202105239
■006m o d
■007cr#unu||||||||
■020 ▼a9798291568903
■035 ▼a(MiAaPQ)AAI32271986
■035 ▼a(MiAaPQ)umichrackham006453
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a004
■1001 ▼aAlharbi, Rahaf Mansour S.
■24510▼aReimagining Privacy and Transparency in Visual Assistance Technologies with Blind People
■260 ▼a[Sl]▼bUniversity of Michigan▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a146 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 87-03, Section: B.
■500 ▼aAdvisor: Brewer, Robin;Schoenebeck, Sarita.
■5021 ▼aThesis (Ph.D.)--University of Michigan, 2025.
■520 ▼aArtificial intelligence (AI) technologies are often advertised as a "solution" to remove accessibility barriers for blind people. At the same time, AI technologies are often designed without consulting blind people, exposing them to privacy harms and errors that are difficult to detect. Through 46 interviews and 16 focus groups with blind communities, this dissertation showcases that frameworks from disability justice and disability studies can guide the development of responsible AI systems. Specifically, this dissertation examines three interconnected projects around visual assistance technologies (VAT), which are real-world applications that blind people use to gain visual access. The first study explores the tensions of using AI to manage privacy concerns in VAT. Findings revealed that blind participants rejected automating privacy decisions, preferring greater control over when and how privacy is managed. Drawing from disability justice principles, this work generates design recommendations to build AI-enabled privacy tools that foster mutual understanding and collaboration. The second study explores how blind people detect and contest errors in VAT. Findings demonstrated that VAT often produces cultural biases and fails to support blind people in making sense of AI. By employing theories from disability studies, this research explains the contours of how VAT often reflects a limited representation of blind people's needs. The third study bridges the first and second projects, investigating how to support blind people in detecting errors in AI tools for privacy management. Findings indicated that transparency in AI-enabled privacy-enhancing technologies should encompass interaction-level cues and broader technical architecture, informing users of specific outputs and the overall system. The primary contributions of this dissertation are: (1) empirical accounts that detail the perspectives of blind communities on existing and emerging AI systems and affirm the central role blind people play in contesting, imagining, and verifying AI technologies, (2) theoretical implications that advance accessibility scholarship by applying disability studies and disability justice activism to reveal how blind communities reshape AI systems, and (3) design implications to expand the boundaries of responsible AI practice and co-create AI systems that honors the expertise of blind communities.
■590 ▼aSchool code: 0127.
■650 4▼aInformation technology
■650 4▼aInformation science
■650 4▼aComputer science
■650 4▼aDisability studies
■653 ▼aLow vision
■653 ▼aPrivacy
■653 ▼aTransparency
■653 ▼aAccessibility
■653 ▼aHuman-computer interaction
■653 ▼aBlind
■690 ▼a0723
■690 ▼a0489
■690 ▼a0984
■690 ▼a0201
■71020▼aUniversity of Michigan▼bInformation.
■7730 ▼tDissertations Abstracts International▼g87-03B.
■790 ▼a0127
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17359945▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
