Cited by 13Ethics and Social Impacts of AIDigital Education and SocietyArtificial Intelligence in Education

Co-designing AI with youth partners: Enabling ideal classroom relationships through a novel AI relational privacy ethical framework

Open access
Chang, Michael Alan, Tissenbaum, Mike, Philip, Thomas M., D’Mello, Sidney K.·2025·Computers and Education: Artificial Intelligence

Abstract

In recent years, the design of AI-based tools for educational spaces have been largely driven by researchers who impart their past expertises, experiences, and perspectives in the design process. While this typically leads to technically feasible designs and are often well-grounded in theories of learning, youth agency is typically limited in this process. In this paper, we argue that designers have a significant ethical responsibility to incorporate youth voices – in particular, their dreams and concerns – into the design of AI tools starting from conception. This need is particularly important as new applications for AI, such as AI-supported collaboration, introduce new surveillance vectors into classroom spaces. Drawing from recent scholarship which advances ethics and relationality in participatory co-design with youth, we introduce a co-design methodology in which youth are supported in imagining expansive technical possibilities for K-12 public schools, grounded within affordances, limitations, and tradeoffs of AI and machine learning techniques . This approach is demonstrated through our Learning Futures Workshop , which brought together 30 historically minoritized youth in conversation with experts in both education and technology. Through detailed case study on the enactment of the workshop, including a thematic analysis of the activities the youth engaged in and their outputs, we identified new, expansive relational possibilities for AI, ethical commitments to support the design, and finally, developed a novel AI Relational Privacy ethical framework that supports the design of new collaborative AI platforms. We conclude by connecting these findings and frameworks to the design of newly enacted AI-based applications and underlying data infrastructures.

Integrity signals

Registry

Crossref · 10.1016/j.caeai.2025.100364

Cites

65 of 65 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

We couldn’t retrieve this paper’s full text, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Cites

Flagged first, then by citations

4 of the 69 references couldn’t be matched to a known paper.

Cited by

Sorted by citations

Authors
Chang, M. A., Tissenbaum, M., Philip, T. M., & D’Mello, S. K.
Affiliations
  • Boston University (US)
  • University of Illinois Urbana-Champaign (US)
  • University of California, Berkeley (US)
  • University of Colorado Boulder (US)
Corresponding author
Chang, Michael Alan
Venue
Computers and Education: Artificial Intelligence
Publisher
Elsevier BV
Published
2025-06
Volume
8
Pages
100364
ISSN
2666-920X (ISSN-L)
OpenAlex
W4406411734
Cited by
12
References
69
Citations by year
2026: 6 · 2025: 6
FWCI
12.55
Type
journal article
Language
en
Open access
Locations
  • Computers and Education Artificial Intelligence (publishedVersion, cc-by-nc-nd)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
National Science Foundation (2019805, 2021159)
Topics
  • Social Sciences › Safety Research › Ethics and Social Impacts of AI
  • Physical Sciences › Computer Science › Information Systems › Digital Education and Society
  • Physical Sciences › Computer Science › Information Systems › Artificial Intelligence in Education
Keywords
youth participation, participatory design, classroom relationships, K-12 education, AI ethics, relationality, data infrastructure

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review