AI evidence extraction

Ethics of AI in Education: Towards a Community-Wide Framework

W. Holmes, Kaśka Porayska‐Pomsta, Ken Holstein, Emma Sutherland, Toby T. Baker, Simon Buckingham Shum et al. · International Journal of Artificial Intelligence in Education · 2021

AI-generated evidence extraction, verified across multiple analytical personas. Not a substitute for the peer-reviewed original.

5/10

Relevance

3/4

Quality (LMQS)

Evidence

1,056

Citations

109.45

FWCI

Top 10%

Impact

This is an AI analysis. Read the peer-reviewed original at the publisher: https://doi.org/10.1007/s40593-021-00239-1

Methodology & findings

Study design

Qualitative survey study.

Sample

N = 17, 2 groups

Primary method

Qualitative thematic analysis of open-ended survey responses. No formal statistical hypothesis testing or quantitative analysis reported. Responses quoted verbatim with bracketed additions/deletions for clarity.

Main result

All 17 respondents demonstrated overall agreement that AIED researchers do not pay sufficient attention to the ethics of AIED. The survey revealed that "most AIED researchers are not trained to tackle the emerging ethical questions" and that "a well-designed framework for engaging with ethics of AIED that combined a multidisciplinary approach and a set of robust guidelines seems vital in this context." Key ethical concerns identified include data ownership and control, privacy, bias and representation, transparency of decisions, pedagogical choices, user agency, equity and fairness, and potential long-term impacts on students during formative periods.

Reports effect sizes.

Research paradigm

Interpretivist/qualitative

Author conclusions

The authors conclude that "While the low numbers of researchers attending the Ethics of AIED workshops and responding to the survey reported in this paper suggest that the AIED community have a low level of interest in the ethics of AIED, the responses to this survey and the other papers in this SI indicate otherwise. Clearly, many AIED researchers do recognise the importance and value of engaging with the ethics of their work." They further state that "a well-designed framework for engaging with ethics of AIED that combined a multidisciplinary approach and a set of robust guidelines seems vital in this context" and emphasize that "the AIED community needs to debate the value and usefulness of developing an ethical framework and practical guidelines, to inform our ongoing research, and to ensure that the AIED tools that we develop and the approaches that we take are, in the widest sense, ethical by design."

Risk of bias

Selection bias: Respondents selected based on citations and reputation, potentially excluding early-career and underrepresented voices; Self-selection bias: Only 17 of 60 invited researchers completed survey; non-respondents may have different views; Social desirability bias: Respondents aware of attribution; may provide responses aligned with perceived ethical norms; Researcher positionality: Survey designed by ethics-focused researchers may frame questions in ways that encourage ethics-positive responses; Selection bias: Researchers were selected primarily by citations and reputation, potentially excluding emerging voices; Response bias: 28% response rate (17/60) may represent self-selection of researchers more interested in ethics; Researcher positionality: Survey authors are embedded in AIED community, potentially influencing how questions were framed and interpreted; Selection bias: Only 17 of 60 invited researchers responded, potentially introducing selection bias favoring those already interested in ethics; Respondent bias: All respondents were identified as leading researchers by citations/reputation, potentially not representing mid- and early-career researchers adequately; Self-selection bias: Those who chose to respond may have stronger views on ethics than non-respondents; Interviewer bias: Although quotes are attributed verbatim, the selection of which quotes to include in the paper introduces potential bias

Limitations

The survey had a low response rate (17 of 60 invited participants, 28%), which the authors acknowledge suggests "a disappointingly low level of interest in the topic from the broader AIED community." Additionally, the authors note that "it may be difficult to make generalizations across the whole of the AIED community" and acknowledge the challenge that "the ethics of AIED are not yet defined sufficiently to make this useful or viable." The paper itself is descriptive and exploratory rather than providing definitive empirical measurements of effect sizes or quantitative outcomes.

Open questions raised

Lack of agreed frameworks and guidelines for AIED ethics
No published AIED ethics guidelines since Aiken & Epstein (2000); minimal explicit AIED ethics focus in literature
Insufficient research on what teachers and students actually want from AIED systems (e.g., around agency and privacy)
Limited attention to pedagogical ethics in AIED (purpose of learning, choice of pedagogy, role of technology relative to teachers)
Missing focus on equity and representativeness; potential for AIED systems to bias against certain demographic groups
Lack of research on human agency and teacher empowerment in AIED systems

Data: Full survey responses will be made available on request; respondents consented to data sharing; "The full survey responses will be made available on request. The respondents all gave consent for this to happen."Extracted from: pdfAgreement 58%

Explore related topics

Responsible AI Use AI Governance Survey