CoChat
Sign Up
Cited by 78Ethics and Social Impacts of AIArtificial Intelligence in EducationGender and Technology in Education

The DOI matches its Crossref record; 53 of the 53 references it cites were found in a registry or an index; 3 of the 49 sources its claims cite were checked for support.

Assessment of AI ethical reflection: the development and validation of the AI ethical reflection scale (AIERS) for university students

Open access
Wang, Ziying, Chai, Ching-Sing, Li, Jiajing, Lee, Vivian Wing Yan·2025·International Journal of Educational Technology in Higher Education
DOI 10.1186/s41239-025-00519-z
Open this paper in CoChat — free account

Abstract

This study developed and validated the AI Ethical Reflection Scale (AIERS), a tool for measuring university students’ ethical reflection on AI in three dimensions: AI ethical awareness, AI critical evaluation, and AI for social good. This study used a sample of 730 university students, and confirmatory factor analysis supported the three-dimensional structure of the AIERS, which demonstrated good internal consistency and construct validity. Additionally, we found evidence of convergent and discriminant validity, as AI ethical reflection was positively yet modestly correlated with AI literacy, indicating that AI ethical reflection is a related yet distinct construct from AI literacy. The study also explored the differences in AI ethical reflection based on gender, academic discipline, and prior AI experience. The results revealed that female students exhibited higher AI ethical awareness than male students. Furthermore, frequent AI users reported significantly greater AI ethical reflection across all dimensions than those who used AI less frequently or had never used it. However, no significant differences were found between Science, Technology, Engineering, Mathematics (STEM) and non-STEM students. Collectively, these findings emphasize the importance of integrating ethical considerations and hands-on AI experiences from diverse perspectives of gender in AI education across disciplines to promote the ethical and responsible use of AI among university students.

Overview

AI snapshot

Written from the paper’s full text.

TL;DR

This study developed and validated the AI Ethical Reflection Scale (AIERS), a 12-item instrument measuring university students’ ethical reflection on AI across three dimensions: AI ethical awareness, AI critical evaluation, and AI for social good. Confirmatory factor analysis with 730 students supported the three-dimensional structure, and the scale showed good reliability and validity, with differences found by gender and prior AI use but not by STEM/non-STEM discipline.

Key findings

  • The three-factor model fit the data well: χ²(51) = 216.76, χ²/df = 4.25, GFI = 0.95, CFI = 0.94, TLI = 0.92, RMSEA = 0.07, SRMR = 0.05, and fit significantly better than a one-factor model (Δχ² = 722.74, Δdf = 3, p < .001).
  • Reliability and validity were supported: Cronbach’s alpha ranged from 0.73 to 0.81, composite reliability from 0.74 to 0.81, AVE from 0.41 to 0.59, and factor loadings from 0.47 to 0.81. AI ethical reflection correlated positively with AI literacy (r = .43, p < .001), with dimension correlations of r = .19 for AI ethical awareness, r = .44 for AI critical evaluation, and r = .39 for AI for social good.
  • Female students reported higher AI ethical awareness than male students (M = 5.78, SD = 0.75 vs. M = 5.62, SD = 0.86; t = 2.59, p = .01), but no significant gender differences appeared for overall AI ethical reflection, AI critical evaluation, or AI for social good.
  • Frequent AI users reported higher overall AI ethical reflection (M = 5.58, SD = 0.61) than less frequent users (M = 5.26, SD = 0.70) and never users (M = 5.19, SD = 0.65), F(2, 727) = 18.28, p < .001, and higher scores across all three dimensions; no significant differences were found between STEM and non-STEM students.

Methods

  • Sample: 730 university students from two universities in mainland China and Hong Kong during the 2023–2024 academic year; 60.7% female, ages 17–30 (M = 21.23, SD = 2.84); 46.4% STEM and 53.6% non-STEM; prior AI use for learning/assessment was 64.0% less frequent, 27.8% frequent, and 8.2% never.
  • Instrument: 12 AIERS items rated on a 7-point Likert scale (1 = strongly disagree to 7 = strongly agree), adapted from prior scales and reviewed by three higher education experts; AI literacy was measured with five items from Laupichler et al. (2023), Cronbach’s alpha = 0.77.
  • Analysis: SPSS 26.0 and AMOS 26.0; normality checks, KMO = 0.85, Bartlett’s test χ²(66) = 2,694.20, p < .001; confirmatory factor analyses comparing one-factor and three-factor models; gender invariance testing; Pearson correlations; independent t-tests and ANOVA with Tukey HSD post hoc tests.

Limitations

  • Stated: the sample was recruited exclusively from university students in a specific Chinese cultural context, limiting generalizability; strict measurement invariance was not achieved at the measurement residual level across genders; gender differences were interpreted using existing literature rather than empirically tested; antecedents and consequences of AI ethical reflection remain largely underexplored.
  • Evident from the text: the study uses a cross-sectional, self-report online questionnaire, so causal claims and links to actual AI behavior are not established; the sample comes from only two universities, which may further restrict representativeness.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1186/s41239-025-00519-z

Cites

53 of 53 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

3 of 49 cited sources checked · 1 supported · 0 flagged · accuracy not yet measured

  • Premisecitation
    § Abstract

    AI education programs have primarily focused on teaching technological knowledge and skills concerning how to use AI for work or academic purposes.

    “the primary focus of these educational programs has been to teach technological knowledge and skills concerning how to use AI for work or academic purposes”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Zhai et al., 2021 source

      The Introduction says that learning about AI has begun to enter school curricula, but does not describe what those programs primarily teach. Searches for technological skills and using AI for work or academic purposes found no passage establishing the claimed focus; the review chiefly examines applications of AI in education. (LLM, conf=0.73)

      Experimental — accuracy not yet measured
  • Premisecitation
    § Abstract

    When individuals follow potentially prejudiced or harmful recommendations from AI, uncritical trust may lead to adverse outcomes such as discrimination and social exclusion.

    “When individuals follow potentially prejudiced or even harmful recommendations from AI, this uncritical trust may lead to adverse outcomes such as discrimination and social exclusion”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites Araujo et al., 2020 source
    • cites Kapania et al. 2022 source

      The Introduction and §5.3 link uncritical acceptance of AI decisions to reduced recognition of bias and potential harm; the Discussion says algorithmic bias can perpetuate inequalities. These passages support the warning about harmful, discriminatory outcomes, but the paper does not specifically identify social exclusion as an outcome of following AI recommendations. (LLM, conf=0.82)

      Experimental — accuracy not yet measured
    • cites Zajko, 2021 source
  • Findingown data
    § Abstract

    AI ethical reflection was positively yet modestly correlated with AI literacy.

    “AI ethical reflection was positively yet modestly correlated with AI literacy”
    • Also stated in: § Results · plain
      “we found a significant positive correlation between the overarching AI ethical reflection factor and AI literacy (r = .43, p < .001).”
    Quote in the paperSubstantial claim
  • Findingown data
    § Abstract

    The AIERS is a tool for measuring university students' ethical reflection on AI in three dimensions: AI ethical awareness, AI critical evaluation, and AI for social good.

    “This study developed and validated the AI Ethical Reflection Scale (AIERS), a tool for measuring university students’ ethical reflection on AI in three dimensions: AI ethical awareness, AI critical evaluation, and AI for social good.”
    Quote in the paperSubstantial claim
  • Findingown data
    § Abstract

    Confirmatory factor analysis supported the three-dimensional structure of the AIERS.

    “confirmatory factor analysis supported the three-dimensional structure of the AIERS”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim. 31
  • It may state two separate claims as one. 30
  • The quote only partly states this claim. 1

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    AI for social good is recognized as a pivotal principle of ethical AI.

    “Recognized as a pivotal principle of ethical AI (Akula & Garibay, 2021; Floridi et al., 2018), AI for social good encompasses the motivation and behavioral tendency to ethically use AI in a manner that promotes positive social outcomes.”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites Akula & Garibay, 2021 source
    • cites Floridi et al., 2018 source

  • It may state two separate claims as one.

    Findingown data
    § Results

    The three-factor model better represented the data than the one-factor model, as shown by a significant improvement in model fit (Δχ2=722.74, Δdf=3, p<.001).

    “The significant improvement in model fit compared with Model 1 (Δχ² = 722.74, Δdf = 3, p < .001) suggests that Model 2 better represented the data.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data
    § Discussion

    The CFA results demonstrated strong evidence of both reliability and validity for the AIERS.

    “The CFA results demonstrated strong evidence of both reliability and validity, suggesting that the AIERS can be a valuable instrument for future empirical investigations and education on ethical reflection in the context of AI.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data
    § Results

    The measurement error model was significant compared with the unconstrained model (Δχ2=51.08, Δdf=27, p=.003), suggesting non-invariance.

    “the measurement error model was significant compared with the unconstrained model, Δχ² = 51.08, Δdf = 27, p = .003, suggesting non-invariance.”
    Quote in the paperSubstantial claim

  • The quote only partly states this claim.

    Findingown data
    § Results

    The three-factor correlated model showed good fit: χ2(51)=216.76, χ2/df=4.25, GFI=0.95, CFI=0.94, TLI=0.92, RMSEA=0.07, and SRMR=0.05.

    “Model 2 showed a significantly improved and identical fit to the data: χ² (51) = 216.76, χ²/df = 4.25, GFI = 0.95, CFI = 0.94, TLI = 0.92, RMSEA = 0.07, and SRMR = 0.05.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data
    § Results

    Students with frequent AI use had higher overall AI ethical reflection (M = 5.58, SD = 0.61) than those with less frequent use (M = 5.26, SD = 0.70) and those who had never used AI (M = 5.19, SD = 0.65), F(2, 727) = 18.28, p < .001.

    “For the composite score, the students with frequent AI use had higher overall AI ethical reflection (M = 5.58, SD = 0.61) than those with less frequent use (M = 5.26, SD = 0.70) and those who had never used AI (M = 5.19, SD = 0.65), F(2, 727) = 18.28, p < .001.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data
    § Results

    AI literacy showed significant correlations with AI ethical awareness (r=.19, p<.001), AI critical evaluation (r=.44, p<.001), and AI for social good (r=.39, p<.001).

    “AI literacy showed significant correlations with the three dimensions of the scale, including AI ethical awareness (r = .19, p < .001), AI critical evaluation (r = .44, p < .001), and AI for social good (r = .39, p < .001).”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data
    § Results

    No significant differences by gender were found for the overall AI ethical reflection score or for the subcomponents AI critical evaluation and AI for social good.

    “The results of independent t-tests (see Table 5) revealed no significant differences by gender for the overall AI ethical reflection score or for the subcomponents of AI critical evaluation and AI for social good.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data
    § Results

    Significant differences were observed for both the overall AI ethical reflection and the three subcomponents based on students' prior engagement with AI.

    “For AI experience, significant differences were observed for both the overall AI ethical reflection and the three subcomponents based on the students’ prior engagement with AI.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Findingown data
    § Results

    Invariance was achieved for the measurement weight model (Δχ2=9.50, Δdf=9, p>.05) and the structural covariance model (Δχ2=14.97, Δdf=15, p>.05).

    “The results indicated that invariance was achieved for both the measurement weight model (Δχ² = 9.50, Δdf = 9, p >.05) and structural covariance model (Δχ² = 14.97, Δdf = 15, p >.05).”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data
    § Discussion

    Participants scored highest in AI ethical awareness, followed by AI critical evaluation, and lowest in AI for social good.

    “Specifically, the participants scored highest in AI ethical awareness, followed by AI critical evaluation, and lowest in AI for social good.”
    Quote in the paperSubstantial claim

  • It may state two separate claims as one.

    Findingown data
    § Results

    Students who frequently used AI reported the highest level of AI for social good (M = 5.26, SD = 1.04), followed by those with less frequent use (M = 4.77, SD = 1.05) and those who had never used AI (M = 4.32, SD = 1.28), F(2, 727) = 23.69, p < .001.

    “Lastly, the students who frequently used AI reported the highest level of AI for social good (M = 5.26, SD = 1.04), followed by those with less frequent use (M = 4.77, SD = 1.05) and those who had never used AI (M = 4.32, SD = 1.28), F(2, 727) = 23.69, p < .001.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    Goldsmith et al. (2020) argued that AI ethical reflection can consist of two primary facets: ethical description and felt ethical responsibility.

    “Goldsmith et al. (2020) argued that AI ethical reflection can consist of two primary facets: ethical description and felt ethical responsibility.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Goldsmith et al., 2020 source

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    AI technologies can be misused, resulting in negative consequences for both the individual and society.

    “AI technologies can be misused, resulting in negative consequences for both the individual and society”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Federspiel et al., 2023 source
    • cites Paraman & Anamalah, 2023 source

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    Researchers have stressed that AI education should go beyond literacy and foster ethical reflection.

    “researchers have stressed that AI education should go beyond literacy and foster ethical reflection”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Goldsmith et al., 2020 source
    • cites Jobin et al., 2019 source

  • Probably not a substantial claim. It may be too vague to check against evidence. It may state two separate claims as one.

    Premisecitation
    § Abstract

    AI literacy in ethics can extend beyond awareness and can involve behavioral tendencies that reflect individuals' critical evaluation of AI and its responsible use for the benefit of society.

    “AI literacy in ethics can extend beyond awareness and can involve behavioral tendencies that reflect individuals’ critical evaluation of AI and its responsible use for the benefit of society”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Chiu et al., 2024 source
    • cites Chiu, Chen, Chiu et al., 2024a, b

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim. It may state two separate claims as one.

    Premisecitation
    § Abstract

    AI outputs are influenced by the data the AI is trained on, and discriminatory and harmful outcomes can result when biased data are used.

    “AI outputs are influenced by the data the AI is trained on, and discriminatory and harmful outcomes can result when biased data are used”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Ray, 2023 source

  • It may state two separate claims as one.

    Study design and conductown data
    § Abstract

    The AIERS demonstrated good internal consistency and construct validity.

    “which demonstrated good internal consistency and construct validity”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. It may state two separate claims as one.

    Interpretationreasoning
    § Abstract

    Moral motivation and moral implementation directly correspond to felt ethical responsibility, and both emphasize the personal drive to act on ethical principles and assume a positive social role.

    “These two components directly correspond to the facet of felt ethical responsibility, and both emphasize the personal drive to act on ethical principles and assume a positive social role.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. It may be too vague to check against evidence. It may state two separate claims as one.

    Premisecitation
    § Abstract

    Individuals must not only acquire basic knowledge to comprehend how AI operates but also develop competencies to reflect on ethical issues related to AI, critically evaluate its outputs, and use it responsibly for social good.

    “individuals must not only acquire the basic knowledge to comprehend how AI operates but also develop the competencies to reflect on the ethical issues related to the application of AI, critically evaluate its outputs, and use it responsibly for social good”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Ng et al., 2021

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    Rest's model of moral functioning hypothesizes four key stages of ethical reasoning and behaviors: moral sensitivity, judgment, motivation, and implementation.

    “It hypothesizes four key stages of ethical reasoning and behaviors: moral sensitivity, judgment, motivation, and implementation.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Rest, 1986

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Premisecitation
    § Abstract

    Individuals' behavioral intentions are fundamentally influenced by their knowledge bases and belief systems.

    “Individuals’ behavioral intentions are fundamentally influenced by their knowledge bases and belief systems”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Fishbein & Ajzen, 2009

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim. It may be too vague to check against evidence.

    Interpretationown data
    § Discussion

    The AIERS can be a valuable instrument for future empirical investigations and education on ethical reflection in the context of AI.

    “The CFA results demonstrated strong evidence of both reliability and validity, suggesting that the AIERS can be a valuable instrument for future empirical investigations and education on ethical reflection in the context of AI.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    AI ethical awareness consists of five major aspects: transparency, fairness, privacy, responsibility, and non-maleficence.

    “There is an increasing consensus that AI ethical awareness consists of five major aspects: transparency, fairness, privacy, responsibility, and non-maleficence”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Jang et al., 2022 source
    • cites Jobin et al., 2019 source

  • Probably not a substantial claim. It may state two separate claims as one.

    Premisecitation
    § Discussion

    Gilligan's (1982) theory of moral development argued that females are more likely to be socialized to prioritize care and relationships, which may give them enhanced sensitivity to ethical issues that impact people.

    “Gilligan’s (1982) theory of moral development argued that females are more likely to be socialized to prioritize care and relationships, which may give them enhanced sensitivity to ethical issues that impact people.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Gilligan, 1982

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim. The paper may give no reason why it holds.

    Interpretationreasoning
    § Method

    Individuals with higher AI literacy may be more capable of critically evaluating AI and recognizing its ethical implications.

    “Individuals with higher AI literacy may be more capable of critically evaluating AI and recognizing its ethical implications.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Interpretationown data
    § Results

    The AIERS captures a unique aspect of AI understanding, namely AI ethical reflection, which is complementary to yet distinct from AI literacy.

    “These results validated the AIERS as a measure that captures a unique aspect of AI understanding, namely AI ethical reflection, which is complementary to yet distinct from AI literacy.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim.

    Interpretationcitation
    § Abstract

    AI for social good reflects the final two components of moral motivation and implementation, emphasizing prioritizing and actualizing ethical considerations.

    “AI for social good reflects the final two components of moral motivation and implementation, emphasizing prioritizing and actualizing ethical considerations.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Study design and conductnone given
    § Method

    Students were informed about the voluntary and autonomous nature of their participation and provided consent.

    “the students were informed about the voluntary and autonomous nature of their participation and provided their consent”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Study design and conductown data
    § Results

    Composite reliability coefficients ranged from 0.74 to 0.81 and AVE values ranged from 0.41 to 0.59, providing support for internal consistency and convergent validity of the AIERS constructs.

    “The composite reliability coefficients ranged from 0.74 to 0.81, all higher than the recommended benchmark of 0.70, and the average variance extracted (AVE) values ranged from 0.41 to 0.59, greater than the minimum threshold of 0.40, providing further support for the good internal consistency and convergent validity of the constructs in the AIERS”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Fornell & Larcker, 1981 source

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    A general awareness of AI ethics can be a prerequisite for responsible AI use.

    “A general awareness of AI ethics can be a prerequisite for responsible AI use”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Tigard, 2021 source

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    AI ethical reflection can be related to a sense of internalized responsibility for the ethical application of AI.

    “This ethical reflection can be related to a sense of internalized responsibility for the ethical application of AI”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Paraman & Anamalah, 2023 source

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    Moral development can differ by gender because of distinct gender-related socialization patterns.

    “Existing literature has suggested that moral development can differ by gender because of distinct gender-related socialization patterns”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Gilligan, 1982

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites You et al., 2011 source

  • Probably not a substantial claim. It may state two separate claims as one.

    Premisecitation
    § Discussion

    Strict invariance at the measurement error level is rarely achieved and can be an overly restrictive test for assessing factorial invariance.

    “According to Byrne (2010), strict invariance at the measurement error level is rarely achieved, and it can be an overly restrictive test for assessing factorial invariance.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Byrne, 2010

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.

    Premisecitation
    § Method

    A 7-point Likert scale provides sufficient points of differentiation without overcomplicating response options, reducing central tendency bias and increasing measurement reliability and validity.

    “A 7-point Likert scale provides sufficient points of differentiation without overcomplicating the response options.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Joshi et al., 2015 source
    • cites Taherdoost, 2019

  • It may state two separate claims as one.

    Study design and conductown data
    § Results

    All three dimensions exhibited satisfactory internal consistency, with Cronbach's alpha coefficients ranging from 0.73 to 0.81, exceeding 0.70.

    “all three dimensions exhibited satisfactory internal consistency, with Cronbach’s alpha coefficients ranging from 0.73 to 0.81, exceeding the cutoff value of 0.70”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Fornell & Larcker, 1981 source

  • Probably not a substantial claim. It may be too vague to check against evidence.

    Interpretationreasoning
    § Abstract

    AI critical evaluation can be considered a vital element of AI ethical reflection.

    “AI critical evaluation can be considered a vital element of AI ethical reflection.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. It may be too vague to check against evidence.

    Interpretationcitation
    § Abstract

    Lack of transparency in AI systems highlights the need for individuals to cultivate competencies in critical evaluation.

    “This lack of transparency highlights the need for individuals to cultivate competencies in critical evaluation”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Many AI systems work as ‘black boxes’, making it difficult for users to learn how their outputs are generated.

  • It may state two separate claims as one.

    Study design and conductown data
    § Method

    Cronbach's alpha for the AI literacy measure in the present study was 0.77, suggesting good measure reliability.

    “Cronbach’s alpha for the present study was 0.77, suggesting good measure reliability.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim. It may state two separate claims as one.

    Premisecitation
    § Method

    AI literacy reflects individuals' understanding of AI technologies and their applications, with ethical considerations being a crucial dimension.

    “AI literacy reflects individuals’ understanding of AI technologies and their applications, with ethical considerations being a crucial dimension”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Chiu et al., 2024a

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Ng et al., 2021

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim. It may be too vague to check against evidence.

    Premisecitation
    § Abstract

    AI for social good embodies individuals' commitment to ensuring that AI plays a socially beneficial role rather than aggravating inequalities or causing harm.

    “It embodies individuals’ commitment to ensuring that AI plays a socially beneficial role rather than aggravating inequalities or causing harm”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Floridi et al., 2018 source

  • Probably not a substantial claim. It may be too vague to check against evidence.

    Interpretationreasoning
    § Abstract

    Critical thinking may be imperative in AI ethical reflection to critically reflect on AI processes and outputs to identify and address potential biases.

    “critical thinking may be imperative in AI ethical reflection, which represents individuals’ capability to critically reflect on the processes and outputs of AI to identify and address potential biases inherent in AI systems.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Many AI systems work as ‘black boxes’, making it difficult for users to learn how their outputs are generated.
    • Rests on an assumption: AI can be affected by algorithmic bias.

  • Probably not a substantial claim. It may state two separate claims as one.

    Premisecitation
    § Abstract

    By acquiring general AI ethical awareness, individuals can better anticipate potential risks of AI and have more positive and ethical engagement with AI.

    “By acquiring this general AI ethical awareness, individuals can better anticipate the potential risks of AI and can have more positive and ethical engagement with AI”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Jang et al., 2022 source

  • It may state two separate claims as one.

    Study design and conductown data
    § Results

    All measured items demonstrated acceptable skewness and kurtosis, suggesting the data were approximately normally distributed.

    “All of the measured items demonstrated acceptable skewness values (ranging from − 1.48 to − 0.44) and kurtosis values (ranging from − 0.03 to 2.62), which were below the maximum threshold of|3|, suggesting that the data for all of the items were approximately normally distributed”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Farooq, 2016 source

  • Probably not a substantial claim.

    Premisecitation
    § Discussion

    Direct experiences can be crucial sources of reflection during the learning process.

    “This is in line with experiential learning theory (Kolb et al., 2001), which posits that direct experiences can be crucial sources of reflection during the learning process.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Kolb et al., 2001

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • It may state two separate claims as one.

    Study design and conductcitation
    § Results

    All 12 items had factor loadings above 0.40, ranging from 0.47 to 0.81, demonstrating adequate representation of their respective constructs.

    “all 12 items had factor loadings above the recommended threshold value of 0.40, ranging from 0.47 to 0.81, demonstrating that they adequately represented their respective constructs”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Byrne, 2010

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim. The paper’s conclusions may not depend on it.

    Study design and conductnone given
    § Method

    The study received approval from the University Ethics Committee.

    “The study received approval from the University Ethics Committee.”
    Quote in the paperNot a substantial claim

  • It may state two separate claims as one.

    Study design and conductown data
    § Results

    Bartlett's test of sphericity was significant and the KMO value was 0.85, indicating the data were well suited for factor analysis.

    “The results revealed that the result for Bartlett’s test of sphericity was significant [χ² (66) = 2,694.20, p < .001], and the KMO value was 0.85, greater than 0.70, indicating that the data were well suited for factor analysis”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Hutcheson & Sofroniou, 1999

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Pallant, 2016

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • It may state two separate claims as one.

    Interpretationown data
    § Discussion

    AI ethical reflection and AI literacy are related yet distinct constructs.

    “suggesting that AI ethical reflection and AI literacy are related yet distinct constructs.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    Previous research has documented significant differences in attitudes toward AI ethics between students who have previous experience with AI education and those who do not.

    “previous research has documented significant differences in attitudes toward AI ethics between students who have previous experience with AI education and those who do not”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Jang et al., 2022 source

  • Probably not a substantial claim.

    Premisecitation
    § Abstract

    Critical thinking is a key factor related to the process of moral judgment.

    “Prior literature has suggested that critical thinking is a key factor related to the process of moral judgment”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Samanci, 2015 source

  • It may state two separate claims as one.

    Study design and conductown data
    § Results

    The square root of the AVE for each dimension exceeded the correlations between that dimension and the other dimensions, providing evidence of good discriminant validity.

    “the square root of the AVE for each dimension exceeded the correlations between that dimension and the other dimensions, providing further evidence of good discriminant validity.”
    Quote in the paperSubstantial claim
52 of 52

  • The paper relies on this without support.

    There has been a comparative lack of education on AI ethics.

    “There has been a comparative lack of education on AI ethics.”

  • The paper relies on this without support.

    Little attention has been given to ethical reflection competence among a more general population.

    “with little attention given to ethical reflection competence among a more general population”

  • The paper relies on this without support.

    Researchers have not yet offered a thorough framework to fully conceptualize the scope of the components of AI ethical reflection.

    “researchers have not yet offered a thorough framework to fully conceptualize the scope of these components”

  • The paper relies on this without support.

    Scarce research has provided a comprehensive model to operationalize and measure students’ ethical considerations and competencies in engagement with AI.

    “scarce research has provided a comprehensive model to operationalize and measure students’ ethical considerations and competencies in engagement with AI”

  • The paper relies on this without support.

    Many AI systems work as ‘black boxes’, making it difficult for users to learn how their outputs are generated.

    “Many AI systems work as ‘black boxes’, which makes it difficult for users to learn how their outputs are generated, potentially posing ethical concerns.”

  • The paper relies on this without support.

    AI can be affected by algorithmic bias.

    “While AI is a powerful tool, it can be affected by algorithmic bias.”
6 of 6

Statistics recomputed

Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.

  • Consistentχ²(66) = 2694.2reported p < .001 · recomputed p < .001

    “χ2 (66) = 2,694.20, p < .001”

  • ConsistentF(2, 727) = 18.28reported p < .001 · recomputed p < .001

    “F(2, 727) = 18.28, p < .001”

  • ConsistentF(6, 732) = 5.19reported p < .01 · recomputed p < .001

    “F(6, 732) = 5.19, p < .01”

  • ConsistentF(2, 727) = 14.1reported p < .001 · recomputed p < .001

    “F(2, 727) = 14.10, p < .001”

  • ConsistentF(2, 727) = 23.69reported p < .001 · recomputed p < .001

    “F(2, 727) = 23.69, p < .001”

5 of 5

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • AI-awareness ANOVA degrees of freedom conflict with the sample and other AI-use analyses

    Minor

    The AI-awareness analysis reports F(6, 732), whose denominator degrees of freedom alone exceed the stated sample of 730. The analyses of the same three AI-use groups elsewhere report F(2, 727).

1 of 1

Citations

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Related

Metadata

Authors
Wang, Z., Chai, C.-S., Li, J., & Lee, V. W. Y.
Affiliations
  • Auburn University (US)
  • Chinese University of Hong Kong (HK)
Corresponding author
Lee, Vivian Wing Yan
ORCID
Venue
International Journal of Educational Technology in Higher Education
Publisher
Springer Science and Business Media LLC
Published
2025-03-31
Volume
22
Issue
1
ISSN
2365-9440 (ISSN-L)
DOI
10.1186/s41239-025-00519-z
OpenAlex
W4408978692
Semantic Scholar
f3a1454575d5bed99285910cbc489c0016103198
Cited by
78
References
50
Citations by year
2026: 54 · 2025: 24
FWCI
20.08
Type
journal article
Language
en
Open access
Licence
Locations
  • International Journal of Educational Technology in Higher Education (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
University Grants Commission - Hong Kong (4171077)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
  • Social Sciences › Education › Artificial Intelligence in Education
  • Social Sciences › Gender Studies › Gender and Technology in Education
Keywords
critical evaluation of AI, AI for social good, AI literacy, university students, gender differences, STEM students, AI ethics education, scale development and validation

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.

A language model drew the claims from the paper on 2026-11-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

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