CoChat
Sign Up
Cited by 124Academic integrity and plagiarismArtificial Intelligence in EducationArtificial Intelligence in Healthcare and Education

Ensuring academic integrity in the age of ChatGPT: Rethinking exam design, assessment strategies, and ethical AI policies in higher education

Open access
Evangelista, Edmund De Leon·2025·Contemporary Educational Technology
DOI 10.30935/cedtech/15775
Open this paper in CoChat — free accountOpen PDF

Abstract

The rapid advancement of artificial intelligence (AI) technologies, particularly OpenAI’s ChatGPT, has significantly impacted higher education institutions (HEIs), offering opportunities and challenges. While these tools enhance personalized learning and content generation, they threaten academic integrity, especially in assessment environments. This study systematically examines the impact of ChatGPT on academic integrity in HEIs, focusing on exam design, assessment strategies, AI detection tools, and policy frameworks. This research draws from current literature and expert recommendations to identify practical approaches for developing assessments that foster critical thinking and deep cognitive engagement, making them less susceptible to AI-generated content. Key areas explored include the creation of complex, analytical exam formats, deploying advanced AI detection software to counter AI-assisted cheating, and formulating institutional policies that promote the ethical use of AI. This comprehensive framework aims to equip educators and administrators with practical strategies to preserve academic standards while harnessing the potential benefits of AI, ensuring the continued validity of assessments in the AI-driven educational landscape.

AI snapshot

Written from the paper’s full text.

TL;DR

This PRISMA systematic review synthesizes 76 articles on how ChatGPT affects academic integrity in higher education and recommends redesigning exams, innovative assessments, AI detection tools, and ethical AI policies/training. It argues traditional fact-recall assessments are vulnerable, while higher-order, personalized, process-oriented, and multimodal assessments are harder for AI to replicate.

Key findings

  • From 753 initial records, 76 articles were included; the review identifies four themes: effective exam design, innovative assessment strategies, advanced AI detection tools, and comprehensive policy/training.
  • Traditional essays and MCQs are described as vulnerable to AI misuse; recommended alternatives include oral exams, project-based learning, performance-based assessments, real-world problem-solving, interactive feedback, multimodal questions, personalization, process-over-product, and creativity. Figure 2 reports project-based learning as most commonly mentioned, followed by peer assessments and oral exams.
  • AI detection tools’ perceived effectiveness rose from 40% in 2020 to 70% in 2024 but remained below 100%; tools like GPTZero, Turnitin’s AI detection, DetectGPT, ZeroGPT, and GPT-2 detector have limitations with paraphrased/edited AI text, false positives, and newer models.
  • Policy/training recommendations include ethical AI usage, educator training, student training, policy development, governance, ethical guidelines, detection tools, assessment design, and monitoring/enforcement; ethical AI usage was the most frequently mentioned component, with student training slightly lower.

Methods

  • Systematic review following PRISMA; searched Scopus, Google Scholar, and JSTOR in July–August 2024 using combinations of terms for ChatGPT/GPT, higher education, exam/assessment, policy/guidelines/proctoring/strategy, academic integrity, and cheating/plagiarism.
  • Inclusion: 2022–present, higher education, ChatGPT’s effect on academic integrity, strategies/best practices, peer-reviewed English articles. Exclusion: non-higher education/K-12, no theoretical/empirical support, non-English, conference proceedings/non-peer-reviewed, non-open access, preprints/editorials/book chapters; literature reviews used as background but excluded from synthesis.
  • Screening: 753 records, 100 duplicates removed, 200 removed for missing keywords, 321 removed for non-HEI level, 27 non-open access, 29 conference/preprint/book chapters, leaving 76 articles; findings synthesized into themes and tables/figures.

Limitations

  • Stated: scope limited primarily to ChatGPT, little coverage of other emerging AI models; based primarily on existing literature, so empirical data on long-term effects in diverse settings are needed; focuses on HEIs and not K-12/vocational training; rapid AI evolution limits long-term relevance.
  • Evident from text: the review relies on secondary sources and does not test the recommended exam designs or detection tools in live classrooms; detection effectiveness is reported as “perceived” and remains below 100%, and tool comparisons note false positives and difficulty with paraphrased/edited AI content.

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

Integrity signals

Registry

Crossref · 10.30935/cedtech/15775

Cites

35 of 35 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 43 cited sources checked · 0 supported · 0 flagged

  • Findingreasoning

    Traditional evaluation formats, particularly those that depend on factual recall, are becoming more vulnerable to AI misuse.

    “Table 2 shows that traditional evaluation formats, particularly those that depend on factual recall, are becoming more vulnerable to AI misuse.”
    Quote in the paperSubstantial claim
  • Findingown data

    Diverse AI detection tools, including text recognition software, AI-enhanced plagiarism detection, and linguistic profiling, are being developed and utilized to identify AI-generated content in student work.

    “As outlined in Table 7, the findings reveal that diverse AI detection tools, including text recognition software, AI-enhanced plagiarism detection, and linguistic profiling, are being developed and utilized to identify AI-generated content in student work.”
    Quote in the paperSubstantial claim
  • Findingreasoning

    Oral exams stimulate critical thinking and deeper cognitive engagement by requiring students to articulate and defend their reasoning in real-time.

    “The findings emphasize that oral exams stimulate critical thinking and deeper cognitive engagement by requiring students to articulate and defend their reasoning in real-time.”
    Quote in the paperSubstantial claim
  • Findingcitation

    Traditional assessments such as essays and multiple-choice questions are vulnerable to AI misuse.

    “For instance, traditional assessments such as essays and multiple-choice questions (MCQs) are seen as vulnerable to AI misuse.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Gamage et al., 2023

      Not checked in the free run.

    • cites Susnjak & McIntosh, 2024

      Not checked in the free run.

  • Findingreasoning

    Constructing exams highlighting real-world problem-solving, interactive feedback mechanisms, and multimodal questions are effective techniques for reducing the misuse of AI in exams.

    “The findings, as shown in Table 4, demonstrate that constructing exams highlighting real-world problemsolving, interactive feedback mechanisms, and multimodal questions are effective techniques for reducing the misuse of AI in exams.”
    Quote in the paperSubstantial claim
  • May combine two claims. 47
  • Classified as probably not a substantial claim. 51
  • The quote only partly states this claim. 1

  • May combine two claims.

    Findingown data

    Innovative assessment strategies with a strong emphasis on personalization, collaboration, and real-time feedback can significantly enhance student engagement and reduce AI misuse.

    “The results in Table 6 reveal that innovative assessment strategies, with a strong emphasis on personalization, collaboration, and most importantly, real-time feedback, can significantly enhance student engagement and reduce AI misuse.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingcitation

    Encouraging students to critique or improve AI-generated content helps them engage more critically with the material and AI itself, deepens their understanding, and builds awareness of AI's limitations, turning AI into a learning tool rather than a source of shortcuts.

    “Encouraging students to critique or improve AI-generated content helps them engage more critically with both the material and AI itself.This deepens their understanding of the subject and builds awareness of AI's limitations, turning AI into a learning tool rather than a source of shortcuts (Dai et al., 2023;Gamage et al., 2023).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Dai et al., 2023

      Not checked in the free run.

    • cites Gamage et al., 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Findingnone given

    Sophisticated algorithms and machine learning models are being deployed to enhance accuracy as AI tools progress, offering stronger protections against academic dishonesty.

    “Furthermore, sophisticated algorithms and machine learning models are being deployed to enhance accuracy as AI tools progress, offering stronger protections against academic dishonesty.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Project-based learning is the most commonly mentioned innovative assessment strategy, followed by peer assessments and oral exams; self-reflective assessments and open-book exams are less frequently mentioned but still regarded as important.

    “Additionally, the graph in Figure 2 demonstrates that project-based learning is the most commonly mentioned innovative assessment strategy, followed by peer assessments and oral exams.Self-reflective assessments and open-book exams are less frequently mentioned, but they are still regarded as important in addressing the challenges posed by AI in education.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingreasoning

    AI detection tools like GPTZero, Turnitin's AI detection feature, and DetectGPT can help identify AI-generated content, but their effectiveness is limited by challenges like detecting paraphrased or edited AI responses and the rapid evolution of AI models.

    “While tools like GPTZero, Turnitin's AI detection feature, and DetectGPT can help identify AI-generated content, their effectiveness is limited by challenges like detecting paraphrased or edited AI responses and the rapid evolution of AI models.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Findingnone given

    Many existing policies are outdated and lack sufficient clarity on the unauthorized use of AI technologies.

    “However, many existing policies are outdated and lack sufficient clarity on the unauthorized use of AI technologies.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingnone given

    Project-based learning, oral exams, and peer assessments are effective in promoting critical thinking and creativity, areas where AI tools like ChatGPT fall short.

    “The findings in this theme demonstrate the effectiveness of project-based learning, oral exams, and peer assessments in promoting critical thinking and creativity, areas where AI tools like ChatGPT fall short.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: AI-generated content is less effective in critical thinking and creativity.

  • May combine two claims.

    Findingcitation

    Low-stakes, continuous formative assessments provide ongoing feedback, allowing students to reflect on their learning progress, reducing reliance on one-time high-stakes exams that AI can easily exploit, promoting consistent engagement and personal growth.

    “Low-stakes, continuous assessments provide ongoing feedback, allowing students to reflect on their learning progress.This reduces reliance on one-time, high-stakes exams that AI can easily exploit, promoting consistent engagement and personal growth (Rudolph et al., 2024;Susnjak & McIntosh, 2024).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Rudolph et al., 2024

      Not checked in the free run.

    • cites Susnjak & McIntosh, 2024

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Findingnone given

    Perceived effectiveness of AI detection tools increased from 40% in 2020 to 70% in 2024.

    “Perceived effectiveness increased from 40% in 2020 to 70% in 2024, showing improvements in these tools' ability to detect AI-generated content.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Current AI detection tools' effectiveness remains below 100%, suggesting they must fully address the complexities of detecting advanced AI outputs and that advancements are needed to keep pace with evolving AI technologies.

    “Despite this upward trend, effectiveness remains below 100%, suggesting that current AI detection tools must fully address the complexities of detecting advanced AI outputs and that advancements are needed to keep pace with evolving AI technologies.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Findingcitation

    Shifting toward alternative formats like oral exams, open-ended questions, and performance-based evaluations requires more profound understanding and personal insights from students, reducing the risk of AI exploitation.

    “To address this, some studies recommend shifting toward alternative formats like oral exams, open-ended questions, and performance-based evaluations, which require more profound understanding and personal insights from students, reducing the risk of AI exploitation (Gamage et al., 2023;Susnjak & McIntosh, 2024).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Gamage et al., 2023

      Not checked in the free run.

    • cites Susnjak & McIntosh, 2024

      Not checked in the free run.

    • Rests on an assumption: AI-generated content is less effective in critical thinking and creativity.

  • Classified as probably not a substantial claim.

    Findingcitation

    Incorporating real-world problem-solving tasks and scenarios challenges students beyond factual recall, which AI can easily handle.

    “Articles also suggest incorporating real-world problem-solving tasks and scenarios to challenge students beyond factual recall, which AI can easily handle (Chaudhry et al., 2023;Rudolph et al., 2024).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Chaudhry et al., 2023

      Not checked in the free run.

    • cites Rudolph et al., 2024

      Not checked in the free run.

  • May combine two claims.

    Findingcitation

    Using practical, real-world scenarios in assessments requires students to apply theoretical knowledge in ways that AI cannot easily replicate, fostering problem-solving, creativity, and critical thinking, preparing students for real-world challenges while minimizing AI misuse.

    “Using practical, real-world scenarios in assessments requires students to apply theoretical knowledge in ways that AI cannot easily replicate.This fosters problem-solving, creativity, and critical thinking, ensuring that students are better prepared for real-world challenges while minimizing the potential for AI misuse (Chaudhry et al., 2023;Moorhouse et al., 2023).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Chaudhry et al., 2023

      Not checked in the free run.

    • cites Moorhouse et al., 2023

      Not checked in the free run.

  • May combine two claims.

    Findingown data

    Effective exam redesign in the AI era involves concentrating on personalization, process-oriented assessments, and fostering creativity.

    “The data in Table 5 illustrates that effective exam redesign in the AI era involves concentrating on personalization, process-oriented assessments, and fostering creativity.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims. The quote only partly states this claim.

    Findingreasoning

    Four major components are essential for responsibly integrating AI into education: ethical AI usage, educator training, student training, and policy development.

    “In addition, this study identified four major components that are essential for responsibly integrating AI into education.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Findingnone given

    Approaches such as project-based learning, oral exams, and open-ended questions have emerged as more effective methods for assessing critical thinking and creativity, which AI cannot easily replicate.

    “Approaches such as project-based learning, oral exams, and open-ended questions have emerged as more effective methods for assessing critical thinking and creativity, which AI cannot easily replicate.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: AI-generated content is less effective in critical thinking and creativity.

  • Classified as probably not a substantial claim.

    Findingreasoning

    The findings emphasize transitioning from traditional fact-based assessments to formats prioritizing higher-order cognitive skills, such as critical thinking, creativity, and real-world problem-solving.

    “In alignment with these objectives, the findings emphasize transitioning from traditional fact-based assessments to formats prioritizing higher-order cognitive skills, such as critical thinking, creativity, and realworld problem-solving.”
    Quote in the paperSubstantial claim

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

    Findingnone given

    The integration of advanced AI detection tools enhances the ability to identify AI-generated content, thereby safeguarding the authenticity of academic submissions.

    “The integration of advanced AI detection tools, as discussed, enhances the ability to identify AI-generated content, thereby safeguarding the authenticity of academic submissions.”
    Quote in the paperNot a substantial claim

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

    Findingreasoning

    These approaches reduce vulnerability to AI misuse while promoting authentic student engagement.

    “These approaches reduce vulnerability to AI misuse while promoting authentic student engagement.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    GPT-3 can be an effective collaborator in paper design.

    “Experimental results revealed that GPT-3 can be an effective collaborator in paper design.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Mcllwraith et al. (2023)

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    Stanford University changed its assessment design from traditional essays to more interactive, project-based evaluations, which demand students apply theoretical knowledge to real-world problems, making it difficult for AI tools such as ChatGPT to generate appropriate responses.

    “For example, Stanford University changed its assessment design from traditional essays to more interactive, project-based evaluations.This method demands students to apply theoretical knowledge to real-world problems, making it difficult for AI tools such as ChatGPT to generate appropriate responses.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Plata et al., 2023

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    Reorienting assessments toward higher levels of Bloom's Taxonomy fosters deeper learning and strengthens academic integrity by making it harder for AI-generated responses to pass undetected.

    “Therefore, reorienting assessments toward these higher levels of Bloom's Taxonomy fosters deeper learning and strengthens academic integrity by making it harder for AI-generated responses to pass undetected.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Traditional assessments often emphasize lower-order cognitive skills.
    • Rests on an assumption: AI-generated content is less effective in critical thinking and creativity.

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

    Premisecitation

    ChatGPT enhances personalized learning, automates routine tasks, and provides students instant feedback, tutoring, and support across various academic disciplines.

    “It enhances personalized learning, automates routine tasks, and provides students instant feedback, tutoring, and support across various academic disciplines (Halaweh, 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Halaweh, 2023

      Not checked in the free run.

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

    Premisecitation

    Students who primarily depend on ChatGPT to finish their homework risk developing a dependency that impedes their intellectual development and hinders their capacity to become interactive, objective, and imaginative.

    “Vargas-Mutillo et al. (2023) observed that students who primarily depend on ChatGPT to finish their homework risk developing a dependency that impedes their intellectual development.This dependency hinders their capacity to become interactive, objective, and imaginative.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Vargas-Mutillo et al. (2023)

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    ChatGPT can behave like a magic wand to accomplish tasks quickly and competently, making it nearly impossible to tell the difference between students' words and those of the AI tool.

    “Klyshbekova and Abbott (2024) revealed that ChatGPT could behave like a magic wand to accomplish tasks quickly and competently, making it nearly impossible to tell the difference between students' words and those of the AI tool.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Klyshbekova and Abbott (2024)

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    GPT-4 can generate high-quality and credible academic content, making it challenging to identify work generated by AI.

    “Moreover, the study of Mapletoft et al. (2024) demonstrated that GPT-4 could generate high-quality and credible academic content, making it challenging to identify work that is generated by AI.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Mapletoft et al. (2024)

      Not checked in the free run.

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

    Study design and conductown data

    Seventy-six articles were selected for inclusion in the literature search.

    “Finally, 76 articles were selected for inclusion in the literature search.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationnone given

    Multimodal assessments, which combine various media and open-ended tasks that emphasize critical thinking, promote deeper cognitive engagement, making it difficult for AI to generate accurate responses.

    “Multimodal assessments, which combine various media and open-ended tasks that emphasize critical thinking, promote deeper cognitive engagement, making it difficult for AI to generate accurate responses.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Interpretationreasoning

    ChatGPT and other AI tools can effortlessly manage MCQs and essay-based assessments, causing educators difficulty upholding academic integrity.

    “Since ChatGPT and other AI tools can effortlessly manage MCQs and essay-based assessments, educators have difficulty upholding academic integrity.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    The University of Sydney implemented oral examinations and group projects in response to the rise of AI tools; these formats foster critical thinking and collaboration, lowering the risk of AI misuse while promoting deeper learning.

    “Equally, the University of Sydney implemented oral examinations and group projects in response to the rise of AI tools.These formats foster critical thinking and collaboration, lowering the risk of AI misuse while promoting deeper learning (Cotton et al., 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Cotton et al., 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationnone given

    Real-time feedback can provide students with immediate insights into their performance, fostering a sense of urgency and continuous improvement.

    “Moreover, real-time feedback, in particular, can provide students with immediate insights into their performance, fostering a sense of urgency and continuous improvement.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Integrating complex multimodal questions with text and visuals bolsters assessments and decreases vulnerability to AI-generated content.

    “Integrating complex multimodal questions with text and visuals bolsters assessments and decreases vulnerability to AI-generated content.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Susnjak and McIntosh, 2024

      Not checked in the free run.

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

    Premisecitation

    MCQs, mainly in university assessments such as professional licensing exams, are crucial when assessing higher-order learning.

    “stating that MCQs, mainly in university assessments such as professional licensing exams, are crucial when assessing higher-order learning.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Newton and Xiromeriti, 2024

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    These formats promote critical thinking, creativity, and quick problem-solving skills that are difficult for AI systems to imitate.

    “These formats promote critical thinking, creativity, and quick problem-solving skills that are difficult for AI systems to imitate.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: AI-generated content is less effective in critical thinking and creativity.

  • Classified as probably not a substantial claim.

    Study design and conductcitation

    The review followed PRISMA guidelines to guarantee a comprehensive and transparent reporting process.

    “The review pursued the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines (Page et al., 2021) to guarantee a comprehensive and transparent reporting process.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Page et al., 2021

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    Innovative assessment strategies stimulate critical thinking, creativity, real-world problem-solving, and personalized learning, nurturing authentic student engagement and minimizing the risk of AI misuse.

    “These strategies stimulate critical thinking, creativity, real-world problem-solving, and personalized learning, ultimately nurturing authentic student engagement and minimizing the risk of AI misuse, thus upholding academic integrity.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    Changing away from traditional exam formats allows educators to ensure students engage more deeply with the material while promoting academic integrity.

    “Changing away from traditional exam formats allows educators to ensure students engage more deeply with the material while promoting academic integrity.”
    Quote in the paperSubstantial claim

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

    Interpretationnone given

    Project-based learning encourages critical thinking by engaging students in complex, real-world scenarios that require applying theoretical knowledge creatively.

    “Project-based learning encourages critical thinking by engaging students in complex, real-world scenarios that require applying theoretical knowledge creatively.”
    Quote in the paperNot a substantial claim

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

    Premisecitation

    When adequately structured, MCQs cover a wide range of topics and can deliver more comprehensive assessments of student comprehension.

    “When adequately structured, these MCQs cover a wide range of topics and can deliver more comprehensive assessments of student comprehension.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Newton and Xiromeriti, 2024

      Not checked in the free run.

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

    Premisecitation

    As AI technology advances, detection tools must become more sophisticated to keep up with AI's growing capabilities in academic settings.

    “These authors argue that as AI technology advances, detection tools must become more sophisticated to keep up with AI's growing capabilities in academic settings.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Rudolph et al., 2024

      Not checked in the free run.

    • cites Richards et al., 2024

      Not checked in the free run.

  • May combine two claims.

    Interpretationreasoning

    These strategies require students to demonstrate original thinking, continuous engagement, and the capability to combine various information types, making it difficult for AI to generate appropriate or correct answers.

    “These strategies require students to demonstrate original thinking, continuous engagement, and the capability to combine various information types, making it difficult for AI to generate appropriate or correct answers.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Interpretationreasoning

    Oral exams ensure that students demonstrate authentic understanding, a crucial metric that AI-generated responses cannot mimic.

    “This methodology ensures that students demonstrate authentic understanding, a crucial metric that AI-generated responses cannot mimic.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Project-based learning, group discussions, and peer reviews serve as critical methods for combating the misuse of AI in education.

    “Project-based learning, group discussions, and peer reviews are a few of its forms and serve as critical methods for combating the misuse of AI in education.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    These policies and educator and student training are essential for upholding academic integrity and effectively incorporating AI technologies such as ChatGPT into the learning environment.

    “These policies and educator and student training are essential for upholding academic integrity and effectively incorporating AI technologies such as ChatGPT into the learning environment.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Sophisticated algorithms and AI-driven detection tools are needed to detect AI-assisted cheating and ensure fair assessments effectively.

    “The studies of Gamage et al. (2023) and Susnjak and McIntosh (2024) emphasize the need for sophisticated algorithms and AI-driven detection tools to detect AI-assisted cheating and ensure fair assessments effectively.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Gamage et al., 2023

      Not checked in the free run.

    • cites Susnjak and McIntosh, 2024

      Not checked in the free run.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Such strategies can encourage students to engage more authentically while maintaining academic integrity.

    “They emphasize that such strategies can encourage students to engage more authentically while maintaining academic integrity.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Gamage et al., 2023

      Not checked in the free run.

  • Classified as probably not a substantial claim.

    Interpretationnone given

    Overall findings highlight the potential and limitations of current strategies for maintaining academic integrity in the face of AI advancements.

    “Overall, these findings highlight the potential and limitations of current strategies for maintaining academic integrity in the face of AI advancements.”
    Quote in the paperSubstantial claim

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

    Interpretationreasoning

    Clear guidelines and robust training programs are critically important to ensure responsible use of AI in educational settings.

    “The findings in the last theme assert the critical importance of clear guidelines and robust training programs to ensure the responsible use of AI in educational settings.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    Detection tools alone may not be sufficient without complementary strategies such as innovative assessment redesign.

    “It implies that detection tools alone may not be sufficient without complementary strategies such as innovative assessment redesign.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Premisecitation

    Such approaches promote authentic student engagement and help maintain academic integrity.

    “Such approaches promote authentic student engagement and help maintain academic integrity.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Chaudhry et al., 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Study design and conductreasoning

    Following PRISMA improves the research's reproducibility and ensures that the methodology is rigorous, thorough, and bias-free.

    “This process improves the research's reproducibility and ensures that the methodology is rigorous, thorough, and bias-free.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Innovative assessment methods stimulate critical thinking and creativity and make it more challenging for students to misuse AI tools like ChatGPT.

    “These methods stimulate critical thinking and creativity and make it more challenging for students to misuse AI tools like ChatGPT.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Chaudhry et al., 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    Innovative assessment strategies promote active student participation and problem-solving, fostering higher-order cognitive abilities that AI tools cannot easily replicate.

    “It promotes active student participation and problemsolving, fostering higher-order cognitive abilities that AI tools cannot easily replicate.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Higher-order cognitive skills, such as critical thinking, creativity, and problem-solving, are more difficult for AI tools like ChatGPT to replicate than factual recall.

  • May combine two claims.

    Premisecitation

    AI detection methods help educators avoid potential misuse, preserving the credibility and fairness of academic evaluations.

    “As AI tools like ChatGPT become more sophisticated, these detection methods help educators avoid potential misuse, preserving the credibility and fairness of academic evaluations (Moorhouse et al., 2023).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Moorhouse et al., 2023

      Not checked in the free run.

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

    Premisecitation

    Policies must undergo regular updates to keep pace with the rapidly evolving challenges AI presents.

    “With this, De Maio (2024) and Bobula (2024) argue that policies must undergo regular updates to keep pace with the rapidly evolving challenges AI presents.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites De Maio, 2024

      Not checked in the free run.

    • cites Bobula, 2024

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Study design and conductreasoning

    The rapid evolution of AI technology limits the study's long-term relevance.

    “Finally, the rapid evolution of AI technology limits the study's long-term relevance.”
    Quote in the paperNot a substantial claim

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

    Interpretationreasoning

    Together, these approaches reinforce the integrity and fairness of assessments in an AI-driven environment.

    “Together, these approaches reinforce the integrity and fairness of assessments in an AI-driven environment.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The 76 articles systematically reviewed in this study provide a comprehensive and representative picture of the challenges and strategies for maintaining academic integrity in the age of ChatGPT.

  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The four themes-effective exam design, innovative assessment strategies, advanced AI detection tools, and comprehensive policy and training-are interconnected in addressing the challenges posed by AI tools such as ChatGPT in education.

    “To put it concisely, the four themes-effective exam design, innovative assessment strategies, advanced AI detection tools, and comprehensive policy and training-are interconnected in addressing the challenges posed by AI tools such as ChatGPT in education.”
    Quote in the paperSubstantial claim

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

    Interpretationnone given

    AI tools like ChatGPT present challenges but also offer significant opportunities to enrich the educational experience through personalized learning and innovative assessment methods.

    “While AI tools like ChatGPT present challenges, they also offer significant opportunities to enrich the educational experience through personalized learning and innovative assessment methods.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationnone given

    Effective exam design and innovative assessments help turn the emphasis away from easily replicable tasks toward those demanding deeper understanding and creativity, lowering the risk of AI exploitation.

    “Effective exam design and innovative assessments help turn the emphasis away from easily replicable tasks toward those demanding deeper understanding and creativity, lowering the risk of AI exploitation.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    The study's scope is limited primarily to ChatGPT, with little coverage of other emerging AI models and technologies that may pose similar or new challenges in higher education.

    “First, the study's scope is limited primarily to ChatGPT, with little coverage of other emerging AI models and technologies that may pose similar or new challenges in higher education.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationnone given

    Advanced detection tools help identify AI-generated content, and comprehensive policies and training ensure that educators and students have the knowledge and guidelines they need to use AI ethically.

    “At the same time, advanced detection tools help identify AI-generated content, and comprehensive policies and training ensure that educators and students have the knowledge and guidelines they need to use AI ethically.”
    Quote in the paperNot a substantial claim

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

    Study design and conductnone given

    The study concentrates on HEIs and does not account for potential differences in how AI tools impact academic integrity in other educational contexts, such as K-12 education or vocational training.

    “Third, the study concentrates on HEIs and does not account for potential differences in how AI tools impact academic integrity in other educational contexts, such as K-12 education or vocational training.”
    Quote in the paperNot a substantial claim

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

    Interpretationreasoning

    By fostering AI literacy and promoting responsible AI use, these initiatives create a culture of academic integrity that is critical in guiding the challenges posed by AI technologies.

    “By fostering AI literacy and promoting responsible AI use, these initiatives create a culture of academic integrity that is critical in guiding the challenges posed by AI technologies.”
    Quote in the paperNot a substantial claim

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

    Interpretationnone given

    The findings of this study, despite their limitations, can still provide valuable insights and strategies for maintaining academic integrity in AI tools and platforms.

    “However, the findings of this study, despite their limitations, can still provide valuable insights and strategies for maintaining academic integrity in AI tools and platforms.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationreasoning

    As AI tools advance, new forms of academic misconduct may emerge, necessitating the evolution of detection tools.

    “As AI tools advance, new forms of academic misconduct may emerge, necessitating the evolution of detection tools.”
    Quote in the paperNot a substantial claim

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

    Interpretationnone given

    AI technologies, particularly ChatGPT, have transformed the educational landscape, presenting opportunities and challenges for HEIs.

    “AI technologies, particularly ChatGPT, have transformed the educational landscape, presenting opportunities and challenges for HEIs.”
    Quote in the paperNot a substantial claim

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

    Interpretationnone given

    The success of a comprehensive policy framework will rely on its adaptability to the rapid evolution of AI technologies and the effectiveness of enforcement mechanisms.

    “While the policy elements provide a robust foundation, the success of such a framework will rely on its adaptability to the rapid evolution of AI technologies and the effectiveness of enforcement mechanisms.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    The study is based primarily on existing literature.

    “Second, the study is based primarily on existing literature, which means that real-world empirical data on the long-term effects of implementing the recommended exam design and AI detection strategies in diverse educational settings is urgently needed.”
    Quote in the paperSubstantial claim

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

    Interpretationreasoning

    By remaining proactive and continuously refining strategies, institutions can ensure that academic credentials remain credible and that assessments reflect students' authentic abilities in this AI-driven era.

    “By doing so, institutions can ensure that academic credentials remain credible and that assessments reflect students' authentic abilities in this AI-driven era.”
    Quote in the paperNot a substantial claim
70 of 70

  • The paper relies on this without support.

    Only a handful of HEIs have developed a comprehensive framework to adapt to ChatGPT's challenges.

    “While the aforementioned studies highlighted specific issues related to ChatGPT's usage, only a handful of HEIs developed a comprehensive framework to adapt to this tool's challenges.”

  • The paper relies on this without support.

    Traditional assessments often emphasize lower-order cognitive skills.

    “Traditional assessments, which often emphasize lower-order cognitive skills, are particularly vulnerable to AIgenerated content”

  • The paper relies on this without support.

    AI-generated content is less effective in critical thinking and creativity.

    “innovative assessment strategies that challenge students' critical thinking and creativity, two areas where AI-generated content is less effective”

  • The paper relies on this without support.

    ChatGPT misuse undermines the educational process and devalues genuine student effort.

    “Even so, ChatGPT can be used to perpetrate academic fraud, and their misuse undermines the educational process and devalues genuine student effort.”

  • The paper relies on this without support.

    The 76 articles systematically reviewed in this study provide a comprehensive and representative picture of the challenges and strategies for maintaining academic integrity in the age of ChatGPT.

    “Frequency of innovative assessment strategies mentioned, based on the findings of the 76 articles systematically reviewed in this study”

  • The paper relies on this without support.

    Higher-order cognitive skills, such as critical thinking, creativity, and problem-solving, are more difficult for AI tools like ChatGPT to replicate than factual recall.

    “fostering higher-order cognitive abilities that AI tools cannot easily replicate.”
6 of 6

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?

  • LLMs support higher-order cognitive functions aligned with analysis, evaluation, and creation in Bloom's Taxonomy, whereas AWE tools focus on lower-order skills related to remembering and understanding linguistic mechanics.

    “Notably, LLMs support higher-order cognitive functions aligned with analysis, evaluation, and creation in Bloom’s Taxonomy, whereas AWE tools focus on lower-order skills related to remembering and understanding linguistic mechanics (Evangelista, 2025; Khlaif et al., 2025).”

    From Reimagining writing assessment for the AI era: a systematic review on balancing AI support and authentic skill growth (2026) · cites this paper as Evangelista, 2025

    The Introduction (page unavailable) gives the Bloom’s Taxonomy categories but says ChatGPT easily replicates recall and basic comprehension, while higher-order tasks pose a greater challenge for AI. The paper does not make the claimed comparison with AWE tools; its account of LLM capabilities runs contrary to the attributed distinction. (LLM, conf=0.87)

    Checked by a reasoning model — may be wrong.

    Experimental — accuracy not yet measured

  • Multimodal composing requiring integration of text, visuals, and audio resists AI synthesis capabilities.

    “Multimodal composing requiring integration of text, visuals, and audio resists AI synthesis capabilities (Evangelista, 2025).”

    From Reimagining writing assessment for the AI era: a systematic review on balancing AI support and authentic skill growth (2026) · cites this paper as Evangelista, 2025

    The Innovative Assessment Strategies section (page unavailable) says multimodal assessments combine media and open-ended tasks, making accurate AI-generated responses difficult. The paper also discusses student submissions integrating text, visuals, and multimedia, but it does not specifically identify audio or establish that composing across those three modes resists AI synthesis. (LLM, conf=0.83)

    Experimental — accuracy not yet measured

  • The dominance of privacy, data protection, and academic integrity within the ethical discourse reflects the reactive nature of current responses.

    “The dominance of privacy, data protection, and academic integrity within the ethical discourse (Holmes et al., 2023;Evangelista, 2025;Gallent-Torres et al., 2023) further reflects the reactive nature of current responses.”

    From Managing Artificial Intelligence Ethics in Higher Education: A Systematic Framework for Issues and Policy Recommendations (2025) · cites this paper as Evangelista, 2025

    The Policy development section says ethical AI usage was the most frequently mentioned component and strongly emphasizes academic integrity. The passages do not establish that privacy and data protection are also dominant concerns. The characterization of current responses as reactive appears to be the citing authors’ interpretation, rather than a finding this citation must independently support. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

  • Other institutions have pursued distributed governance models, in which departments, faculties, or individual instructors develop localized guidelines tailored to discipline-specific needs.

    “In contrast, other institutions have pursued distributed governance models, in which departments, faculties, or individual instructors develop localized guidelines tailored to disciplinespecific needs (Evangelista, 2025;Grieve et al., 2024).”

    From Managing Artificial Intelligence Ethics in Higher Education: A Systematic Framework for Issues and Policy Recommendations (2025) · cites this paper as Evangelista, 2025

    The Integration of Recommendations Across Educational Settings section discusses discipline-specific assessment approaches and says institutions can develop guidelines for remote assessments (page unavailable). It does not say that institutions have pursued distributed governance or that departments, faculties, or instructors develop localized guidelines. (LLM, conf=0.75)

    Experimental — accuracy not yet measured

  • Generative AI systems are capable of producing original content, providing advanced feedback, and executing cognitive tasks traditionally associated with human expertise.

    “Generative AI (GenAI) systems are capable of producing original content, providing advanced feedback, and executing cognitive tasks traditionally associated with human expertise (Evangelista, 2025).”

    From Managing Artificial Intelligence Ethics in Higher Education: A Systematic Framework for Issues and Policy Recommendations (2025) · cites this paper as Evangelista, 2025

    The Introduction (page not provided) says ChatGPT generates human-like text, essays, and problem solutions and provides instant feedback. The Future Directions table also describes AI delivering adaptive feedback. But the Introduction distinguishes routine cognitive tasks from higher-order work, which it says is harder for AI to mimic; the source does not support the unqualified claim about tasks associated with human expertise. (LLM, conf=0.85)

    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.

  • Chinengundu citation year differs from the reference

    Minor

    The policy framework cites Chinengundu (2023), whereas the only Chinengundu entry in the reference list is dated 2024.

  • Rudolph citation year differs from the reference

    Minor

    The paper repeatedly cites Rudolph et al. as a 2024 source, including in its assessment findings, but the only Rudolph entry in the reference list is dated 2023.

2 of 2

Cites

Flagged first, then by citations

Research works printed but not linked: 3

  1. Chinengundu, T. (2024). Higher education assessment at crossroads on the use of ChatGPT. Proceedings of the International Conference on Future of Education, 6(1), 87–98.
  2. De Maio, C. (2024). Institutional responses to ChatGPT: Analysing the academic integrity policies of four public and private institutions of higher education in Australia. Journal of Academic Language and Learning, 18(1), T1–T8.
  3. Güner, H., Er, E., Akçapinar, G., & Khalil, M. (2024). From chalkboards to AI-powered learning: Students’ attitudes and perspectives on use of ChatGPT in educational settings. Educational Technology & Society, 27(2), 386–404.

Cited by

Sorted by citations

Authors
Evangelista, E. D. L.
Affiliations
Zayed University (AE)
Corresponding author
Evangelista, Edmund De Leon
ORCID
Evangelista, Edmund De Leon
Venue
Contemporary Educational Technology
Publisher
Bastas Publications
Published
2025-01-01
Volume
17
Issue
1
Pages
ep559
ISSN
1309-517X (ISSN-L)
DOI
10.30935/cedtech/15775
OpenAlex
W4405922642
Cited by
124
References
35
Citations by year
2026: 86 · 2025: 38
FWCI
51.54
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
Contemporary Educational Technology (publishedVersion, cc-by)
Topics
  • Social Sciences › Safety Research › Academic integrity and plagiarism
  • Physical Sciences › Computer Science › Information Systems › Artificial Intelligence in Education
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
Keywords
academic integrity, ChatGPT, AI-generated content, test design, assessment methods, AI detection tools, AI ethics policy, higher education, cheating, critical thinking, assessment validity

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 2026-10-06.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-06.

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