CoChat
Sign Up
Cited by 391Artificial Intelligence in EducationArtificial Intelligence in Healthcare and EducationAcademic integrity and plagiarism

Not quite eye to A.I.: student and teacher perspectives on the use of generative artificial intelligence in the writing process

Open access
Barrett, Alex, Pack, Austin·2023·International Journal of Educational Technology in Higher Education
DOI 10.1186/s41239-023-00427-0
Open this paper in CoChat — free accountOpen PDF

Abstract

Generative artificial intelligence (GenAI) can be used to author academic texts at a similar level to what humans are capable of, causing concern about its misuse in education. Addressing the role of GenAI in teaching and learning has become an urgent task. This study reports the results of a survey comparing educators’ (n = 68) and university students’ (n = 158) perceptions on the appropriate use of GenAI in the writing process. The survey included representations of user prompts and output from ChatGPT, a GenAI chatbot, for each of six tasks of the writing process (brainstorming, outlining, writing, revising, feedback, and evaluating). Survey respondents were asked to differentiate between various uses of GenAI for these tasks, which were divided between student and teacher use. Results indicate minor disagreement between students and teachers on acceptable use of GenAI tools in the writing process, as well as classroom and institutional-level lack of preparedness for GenAI. These results imply the need for explicit guidelines and teacher professional development on the use of GenAI in educational contexts. This study can contribute to evidence-based guidelines on the integration of GenAI in teaching and learning.

AI snapshot

Written from the paper’s full text.

TL;DR

This survey compares 158 university students’ and 68 teachers’ perceptions of acceptable generative AI use across six writing-process tasks. Both groups broadly agreed on acceptable and unacceptable uses, but students were generally more accepting of undisclosed AI use while teachers were more accepting of AI for feedback/evaluation, and both reported little institutional policy or training.

Key findings

  • Across tasks, both groups viewed GenAI as more acceptable for early writing stages such as brainstorming and outlining, and for supportive uses like generating ideas or model answers, than for completing writing assignments. Using GenAI to complete writing tasks with or without disclosure was rated unacceptable by most respondents.
  • Significant student–teacher differences appeared mainly for disclosure-related uses: students were more accepting than teachers of undisclosed GenAI use for brainstorming (p = 0.032), outlining (division D: p < 0.001), writing (division D: p < 0.001), and revision (division D: p < 0.001). Teachers were more accepting than students of using GenAI-generated feedback as a model (U = 6397, r = −0.157, p < 0.05) and AI-generated evaluation as a model (U = 6247, r = −0.133, p < 0.05).
  • Preparedness was low: 94.1% of teachers reported their university had no AI policy, 89.7% of teachers had never educated students on acceptable AI use, 95.6% of teachers reported no institutional training on AI, and around 95% of both groups reported no training on AI use.
  • Principal component analysis yielded three factors accounting for 60% of variance: Factor 1 captured highly unacceptable uses (all division D plus writing, outlining, and revising division C), Factor 2 captured generally permissible student-oriented uses (divisions A and B for outlining, revision, brainstorming, and writing), and Factor 3 captured teacher-oriented evaluation and feedback uses.

Methods

  • Cross-sectional questionnaire design with 226 participants: 158 students and 68 teachers. Students were recruited by non-probability voluntary response from a U.S. public research university with course credit; teachers were recruited by purposeful sampling and snowballing across multiple institutions worldwide, primarily second-language teachers. Fifteen incomplete questionnaires were discarded.
  • The instrument presented ChatGPT (GPT-3.5 turbo, February 2023) prompts and outputs for six writing-process tasks—brainstorming, outlining, writing, revising, feedback, and evaluating—and asked respondents to rate four divisions of use (A–D) on a five-point Likert scale. Internal consistency alpha coefficients were above 0.7.
  • Analysis included response frequencies, means and standard deviations, Mann-Whitney U tests for student–teacher comparisons, exploratory factor analysis for survey items, and principal component analysis across writing-process steps and divisions of use. Data were collected in spring 2023 via Qualtrics.

Limitations

  • The paper states that the non-random sample inhibits generalizing findings to a wider population, the heterogeneous teacher and student populations make between-group comparisons tenuous, and the cross-sectional design does not account for the dynamic nature of perceptions. It also notes that GenAI tools and examples may become anachronistic and that the prompts were rudimentary/layperson-level.
  • Evident from the text: the student sample came from a single U.S. public research university and was mostly undergraduate (153/158) and English-native (151/158), while the teacher sample was recruited through snowballing across institutions and was primarily second-language educators (55/68 ENL, 10/68 modern languages), further limiting comparability and generalizability.

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

Integrity signals

Registry

Crossref · 10.1186/s41239-023-00427-0

Cites

45 of 45 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 11 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Results indicate minor disagreement between students and teachers on acceptable use of GenAI tools in the writing process.

    “Results indicate minor disagreement between students and teachers on acceptable use of GenAI tools in the writing process, as well as classroom and institutional-level lack of preparedness for GenAI.”
    Quote in the paperSubstantial claim
  • Findingown data

    There is classroom and institutional-level lack of preparedness for GenAI.

    “Results indicate minor disagreement between students and teachers on acceptable use of GenAI tools in the writing process, as well as classroom and institutional-level lack of preparedness for GenAI.”
    Quote in the paperSubstantial claim
  • Findingown data

    Heavy disagreement was reported by both groups in using GenAI for providing writing feedback without disclosing the use of GenAI.

    “Heavy disagreement was reported by both groups in using GenAI for providing writing feedback without disclosing the use of GenAI.”
    Quote in the paperNot a substantial claim
  • Findingown data

    There was a larger difference between teachers and students when asked if it was ok to use GenAI to brainstorm ideas without disclosing the use of GenAI.

    “However, there was a larger difference between teachers and students when asked if it was ok to use GenAI to brainstorm ideas without disclosing the use of GenAI.”
    Quote in the paperNot a substantial claim
  • Findingown data

    Teachers tended to disagree less than students about using GenAI to model feedback or for evaluation of student writing.

    “As to differences in perspectives, students tended to disagree less than teachers that using GenAI without disclosure was appropriate, and teachers tended to disagree less than students about using GenAI to model feedback or for evaluation of student writing.”
    Quote in the paperNot a substantial claim
  • May combine two claims. 27
  • Classified as probably not a substantial claim. 12

  • May combine two claims.

    Findingown data

    Factor 1 included all of division D as well as writing, outlining, and revising for division C, and these use examples all had the lowest mean agreement of acceptability so we can label this factor as highly unacceptable use.

    “Factor 1 included all of division D as well as writing, outlining, and revising for division C.These use examples all had the lowest mean agreement of acceptability (refer to Table 5) so we can label this factor as highly unacceptable use.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Non-significant differences in mean response frequencies were found along divisions A, C, and D for feedback, however, teachers agreed significantly more than students that division B was acceptable (using GenAI generated feedback as a model), although both groups were generally accepting of this behavior.

    “Non-significant differences in mean response frequencies were found along divisions A, C, and D, however, teachers agreed significantly more than students that division B was acceptable (using GenAI generated feedback as a model) (U = 6397, r = -0.157,p < 0.05), although both groups were generally accepting of this behavior.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Factor 3 contained evaluation and feedback for all A, B and C divisions of use, and the ratings for these use examples were mostly supportive, especially along the A and B divisions of use.

    “Factor 3 contained evaluation and feedback for all A, B and C divisions of use.The ratings for these use examples were mostly supportive, especially along the A and B divisions of use.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Similar to the feedback task, significant differences on perceptions of acceptable use were only detected in division B (using the AI-generated evaluation as a model), again with teachers being more accepting of this use.

    “Similar to the feedback task, significant differences on perceptions of acceptable use were only detected in division B (using the AI-generated evaluation as a model), again with teachers being more accepting of this use (U = 6247, r = -0.133,p < 0.05).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Means for division A of GenAI use (acceptable if the student is already a competent outliner) were not significantly different between groups, but for divisions B, C, and D significant differences were detected with students being more accepting of these uses.

    “Means for division A of GenAI use (acceptable if the student is already a competent outliner) were not significantly different between groups, but for divisions B, C, and D significant differences were detected with students being more accepting of these uses.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Students and teachers both generally felt that using GenAI to brainstorm ideas was acceptable if the student was already a competent brainstormer or only used the output as a model.

    “Students and teachers both generally felt that using GenAI to brainstorm ideas was acceptable if the student was already a competent brainstormer or only used the output as a model.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    There was a tendency to agree that AI would be useful to students and teachers in education, although some trepidation can be insinuated from relatively less agreement (and neutrality for students) to item 3, about AI having a positive impact on education.

    “There was a tendency to agree that AI would be useful to students and teachers in education, although some trepidation can be insinuated from relatively less agreement (and neutrality for students) to item 3, about AI having a positive impact on education.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Concern for student use of AI was fairly high for both teachers and students, with slightly less concern for teacher use.

    “Concern for student use of AI (items 4 and 5) was fairly high for both teachers and students, with slightly less concern for teacher use.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Significant disagreement was found between teachers and students regarding submitting a GenAI-written essay with disclosure and without disclosure, with teachers more heavily disagreeing with this behavior.

    “Significant disagreement was found between teachers and students regarding submitting a GenAI-written essay with disclosure (U = 4311.5,r = -0.169,p = 0.011) and without disclosure (U = 4245.5,r = -0.223,p < 0.001), with teachers more heavily disagreeing with this behavior.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Factor 1 contains the divisions of use that involve GenAI doing all the writing of a writing assignment, which was rated as the most unacceptable use of GenAI.

    “Factor 1 contains the divisions of use that involve GenAI doing all the writing of a writing assignment, which was rated as the most unacceptable use of GenAI.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Around 95% of both groups reported that they had received no training on the use of AI, and about 90% of teachers have not taught their students about the appropriate use of AI.

    “Around 95% of both groups reported that they had received no training on the use of AI, and about 90% of teachers have not taught their students about the appropriate use of AI.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Submitting AI-brainstormed ideas in class was seen as acceptable by half of teachers and students who took the survey, with another 10 to 16% uncertain and the remaining against.

    “Submitting AI-brainstormed ideas in class was seen as acceptable by half of teachers and students who took the survey, with another 10 to 16% uncertain and the remaining against.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Mean familiarity scores were not high, but students reported slightly more familiarity with AI than did teachers, however the difference was non-significant.

    “Mean familiarity scores were not high, but students reported slightly more familiarity with AI than did teachers, however the difference was non-significant.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Although teachers and students were predominantly in alignment in their disagreement that using GenAI to brainstorm without disclosure was acceptable, fewer students disagreed, and about 11% were uncertain.

    “Although teachers and students were predominantly in alignment in their disagreement that this use was acceptable, fewer students disagreed, and about 11% were uncertain.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Teachers and students had comparable perceptions on acceptable use of GenAI for writing an essay if the student was already competent in writing an essay or if a student wanted to use a GenAI draft of an essay to model their own writing.

    “Teachers and students had comparable perceptions on acceptable use of GenAI for writing an essay (Fig. 3) if the student was already competent in writing an essay (Division A pair) or if a student wanted to use a GenAI draft of an essay to model their own writing (Division B pair).”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    Factor 2 contained the most acceptable uses of GenAI across four student-oriented steps of the writing process, which included utilizing GenAI for tasks that the user is already proficient in, and to generate ideas or model answers.

    “Factor 2 contained the most acceptable uses of GenAI across four student-oriented steps of the writing process, which included utilizing GenAI for tasks that the user is already proficient in, and to generate ideas or model answers.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Factor 2 primarily contained divisions of use A and B for outlining, revision, brainstorming, and writing, and these use examples had relatively high means and were seen as generally permissible uses of GenAI, despite writing A having a majority disagreement response frequency.

    “Factor 2, on the other hand, primarily contained divisions of use A and B for outlining, revision, brainstorming, and writing.These use examples had relatively high means and were seen as generally permissible uses of GenAI, despite writing A having a majority disagreement response frequency.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Having GenAI revise an essay showed mixed perceptions between teachers and students with no differences detected for divisions A and B, but significant differences found in divisions C and D.

    “Similar to writing, having GenAI revise an essay (Fig. 4) showed mixed perceptions between teachers and students with no differences detected for divisions A and B, but significant differences found in divisions C and D.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Premisecitation

    In the education sector there is serious concern about academic integrity and plagiarism regarding GenAI.

    “The public discourse surrounding GenAI has been mostly positive, but in the education sector there is serious concern about academic integrity and plagiarism (Dehouche, 2021;Lampropoulos et al., 2023;Sullivan et al., 2023;Yeo, 2023).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Dehouche, 2021 source

      Not checked in the free run.

    • cites Lampropoulos et al., 2023 source

      Not checked in the free run.

    • cites Sullivan et al., 2023 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Student perceptions of academic dishonesty have historically been unclear or incomprehensive, and rarely align with teacher expectations.

    “Student perceptions of academic dishonesty have historically been unclear or incomprehensive, and rarely align with teacher expectations (Tatum, 2022), GenAI will only serve to complicate this (Farrokhnia et al., 2023).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Tatum, 2022 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    There is no identifiable other whose work is being stolen by a chatbot.

    “The result is that there is no identifiable other whose work is being stolen by a chatbot.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Generative AI is trained on a selection of diverse natural language data from across the Internet which allows it to string together unique combinations of words and phrases, similar to how humans learn to produce an unlimited amount of novel spoken or written text from the limited language they absorb from their environment.

    “Generative AI is trained on a selection of diverse natural language data from across the Internet which allows it to string together unique combinations of words and phrases, similar to how humans learn to produce an unlimited amount of novel spoken or written text from the limited language they absorb from their environment, a tenet of generative grammar (Chomsky, 1991).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Chomsky, 1991

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    According to Seow (2002), process writing includes four basic stages: planning (including brainstorming and outlining), writing, and revising or editing.

    “According to Seow (2002), process writing includes four basic stages: planning (including brainstorming and outlining), writing, and revising or editing.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Seow (2002) source

      Not checked in the free run.

  • May combine two claims.

    Study design and conductreasoning

    A key assumption of the Mann-Whitney U test is independent observations from compared groups, and this assumption was satisfied.

    “A key assumption of the Mann-Whitney U test is independent observations from compared groups.This assumption was satisfied as participants completed the questionnaire individually and cross-group contamination was unlikely due to separate questionnaire hyperlinks.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    The seemingly rudimentary prompts exemplified in this study represent how a teacher or student without much experience or training in prompt engineering would likely interact with a GenAI tool like ChatGPT.

    “Regardless, the seemingly rudimentary prompts exemplified in this study represent how a teacher or student without much experience or training in prompt engineering would likely interact with a GenAI tool like ChatGPT.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductown data

    The KMO test value was 0.674, showing an adequate proportion of variance in the survey which could indicate underlying factors.

    “The KMO test value was 0.674, showing an adequate proportion of variance in the survey which could indicate underlying factors.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    The KMI measure of sampling adequacy value was 0.862 and Bartlett's Test of Sphericity was significant (p < 0.001).

    “The KMI measure of sampling adequacy value was 0.862 and Bartlett's Test of Sphericity was significant (p < 0.001).”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    The low reliability of factors 2 and 3 is likely due to only having two items per factor and also disparate attitudes concerning teachers and students from the surveyed groups.

    “The low reliability of factors 2 and 3 is likely due to only having two items per factor and also disparate attitudes concerning teachers and students from the surveyed groups.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    15 incomplete questionnaires were discarded following the collection period.

    “15 incomplete questionnaires were discarded following the collection period.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The heterogeneous population of teachers and students renders between-group comparisons tenuous.

    “Limitations of the study include the non-random sample which inhibits generalizing findings to a wider population, and the heterogeneous population of teachers and students that render between-group comparisons tenuous.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning

    If educators are transparent and clearly state and model acceptable uses of GenAI then students may follow.

    “If educators are transparent and clearly state and model acceptable uses of GenAI then students may follow.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The cross-sectional design of the current study does not take into account the dynamic nature of student and teacher perceptions.

    “Also, the cross-sectional design of the current study does not take into account the dynamic nature of student and teacher perceptions.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    Informed consent from all participants was acquired and all ethical procedures were adhered to according to the standards of the university institutional review board.

    “Informed consent from all participants was acquired and all ethical procedures were adhered to according to the standards of the university institutional review board.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationown data

    The factor 3 use examples were categorized as being teacher-oriented, and their inclusion in one factor here demonstrates that participants conceptualized acceptable teacher use of GenAI differently.

    “The factor 3 use examples were categorized as being teacher-oriented, and their inclusion in one factor here demonstrates that participants conceptualized acceptable teacher use of GenAI differently.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Interpretationreasoning

    By the time this paper is published the GenAI tools and examples of use in this study will likely be anachronistic to a wider variety of available programs that users might interact with.

    “By the time this paper is published the GenAI tools and examples of use in this study will likely be anachronistic to a wider variety of available programs that users might interact with; a common limitation in emerging technology research.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationcitation

    Many teachers may be hesitant to embrace GenAI tools, such as ChatGPT, due to concerns related to cost, privacy, and legality, in addition to a naivety as to how these tools can be appropriately used for educational purposes.

    “Many may be hesitant to embrace GenAI tools, such as ChatGPT, due to concerns related to cost, privacy, and legality (Kumar, 2023), in addition to a naivety as to how these tools can be appropriately used for educational purposes.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Kumar, 2023 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    The measured perceptions herein will likely be useful when considering approaches to GenAI integration in education.

    “For this reason, the measured perceptions herein will likely be useful when considering approaches to GenAI integration in education.”
    Quote in the paperNot a substantial claim
37 of 37

  • The paper relies on this without support.

    No consensus has arisen regarding what constitutes appropriate use of GenAI in higher education.

    “To date no consensus has arisen regarding what constitutes appropriate use of GenAI in higher education.”

  • The paper relies on this without support.

    The usefulness of GenAI for assisting in the writing process will likely lead to its adoption by teachers and students regardless of its limitations.

    “The use of GenAI for education involves other risks such as functional opacity, data privacy, and reliability (Yu & Guo, 2023) which need to be accounted for, but the usefulness of the technology for assisting in the writing process will likely lead to its adoption by teachers and students regardless of these limitations.”

  • The paper relies on this without support.

    The student population is homogeneous and typical of public university undergraduates.

    “The student population was chosen for its homogeneity and typicality of public university undergraduates.”

  • The paper relies on this without support.

    Second language teachers' insights are useful for addressing plagiarism and academic dishonesty.

    “Second language teachers that teach a language different than the students' native language were the primary target of the teacher survey due to their experience with groundup writing instruction and the corresponding need to address plagiarism and academic dishonesty regularly.”

  • The authors state this as an assumption.

    Both student and teacher perceptions of the ethics of using GenAI in writing education will change as individuals become more aware of and competent in the technology, and as the technology develops and changes over time.

    “It is reasonable to assume that both student and teacher perceptions of the ethics of using GenAI in writing education will change as individuals become more aware of and competent in the technology, and as the technology develops and changes over time.”
5 of 5

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?

  • Participants wanted explicit guidelines and professional development on the integration of GenAI in the tertiary educational context.

    “Barrett and Pack (2023) found that participants wanted explicit guidelines and professional development on the integration of GenAI in the tertiary educational context.”

    From A systematic literature review of attitudes, intentions and behaviours of teaching academics pertaining to AI and generative AI (GenAI) in higher education: An analysis of GenAI adoption using the UTAUT framework (2024) · cites this paper as Barrett and Pack, 2023

    The Abstract and Discussion say the survey findings imply a need for explicit guidelines and teacher professional development in educational contexts. The Survey results report that participants lacked AI policies and training, but do not say participants expressed a desire for either. The claim turns the authors’ conclusion into a reported participant preference. (LLM, conf=0.88)

    Experimental — accuracy not yet measured

  • This fear might be a consequence of the universities' initial ban and lack of guidance on GenAI use.

    “This fear might be a consequence of the universities ' initial ban and lack of guidance on GenAI use (Barrett &amp; Pack, 2023; Chan, 2023a; Lim et al., 2023).”

    From Artificial Intelligence for Academic Purposes (AIAP): Integrating AI literacy into an EAP module (2025) · cites this paper as Barrett & Pack, 2023

    The Discussion (page unavailable) reports apprehension about AI and says clear policies and education on acceptable use can address it. The Introduction (page unavailable) notes that some schools banned GenAI, but the paper does not suggest those bans caused the fear; that part of the proposed explanation is unsupported. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • Unguided AI use may lead to superficial engagement or overreliance on generated output.

    “However, unguided AI use may lead to superficial engagement or overreliance on generated output (Barrett & Pack, 2023).”

    From Generative AI, visual scaffolds, and metacognition in EFL writing (2026) · cites this paper as Barrett & Pack, 2023

    The Related literature section (page unavailable) says that students outsourcing writing to AI will likely harm their cognitive ability, which supports concern about reliance on generated work. The Abstract and Discussion call for explicit guidelines, but the paper does not establish that a lack of guidance leads to superficial engagement or overreliance. (LLM, conf=0.82)

    Experimental — accuracy not yet measured

  • In a sample of 68 instructors, 63 of which were employed by higher education, nearly 95% reported that their institutions currently lacked an AI policy.

    “Another survey by Barrett and Pack [15] showed a similar need for AI policy development-in a sample of 68 instructors, 63 of which were employed by HE, nearly 95% reported that their institutions currently lacked an AI policy.”

    From "I'd Never Actually Realized How Big An Impact It Had Until Now": Perspectives of University Students with Disabilities on Generative Artificial Intelligence (2025) · cites this paper as [15]

    The Participants section reports 68 teachers, and the Discussion reports that 94.1% of teachers said their university had no AI-use policy. The passages in our copy do not establish the additional claim that 63 of the 68 teachers were employed in higher education. (LLM, conf=0.87)

    Experimental — accuracy not yet measured

  • A recurring theme in our work was variability in presence and clarity of course AI policies, in line with [15,45].

    “In line with fi ndings from Barrett and Pack [15] and Ghimire and Edwards [45], a recurring theme in our work was the variability in the presence and clarity of course AI policies.”

    From "I'd Never Actually Realized How Big An Impact It Had Until Now": Perspectives of University Students with Disabilities on Generative Artificial Intelligence (2025) · cites this paper as [15]

    The Survey results report mixed student answers about whether their university had an AI policy, while teachers from various institutions generally reported that theirs did not. The Discussion calls for clear AI policies. These passages support concern about policy presence and clarity, but they do not examine variability in course-level policies specifically. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

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.

  • Discussion says competent users accepted AI offloading for every writing step

    Major

    The discussion says this use was generally accepted for all steps, but the results explicitly say both groups predominantly considered it inappropriate for essay writing.

1 of 1

Cites

Flagged first, then by citations

Research works printed but not linked: 3

  1. Chomsky, N. (1991). Linguistics and cognitive science: Problems and mysteries. In Aka Kasher (Ed.), The Chomskyan Turn. Blackwell. pp. 26–53.
  2. Ely, J. J., Henderson, L., & Wachsman, Y. (2013). Testing the effectiveness of the university honor code. Academy of Educational Leadership Journal, 17(4), 95–104.
  3. Jackson, M. C. (2021). Artificial intelligence and algorithmic bias: The issue with technology reflecting history and humans. Journal of Business & Technology Law, 16(2), 299–316.

Cited by

Sorted by citations

Related

Authors
Barrett, A., & Pack, A.
Affiliations
  • Florida State University (US)
  • Brigham Young University–Hawaii (US)
Corresponding author
Barrett, Alex
ORCID
Venue
International Journal of Educational Technology in Higher Education
Publisher
Springer Science and Business Media LLC
Published
2023-11-10
Volume
20
Issue
1
ISSN
2365-9440 (ISSN-L)
DOI
10.1186/s41239-023-00427-0
OpenAlex
W4388551749
Cited by
391
References
45
Citations by year
2026: 147 · 2025: 176 · 2024: 68
FWCI
91.60
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)
Topics
  • Physical Sciences › Computer Science › Information Systems › Artificial Intelligence in Education
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Social Sciences › Safety Research › Academic integrity and plagiarism
Keywords
generative AI, academic writing, writing process, student perceptions, teacher perceptions, ChatGPT, brainstorming, outlining, revision, peer feedback, institutional readiness, teacher professional development

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.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-08.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-08.

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