The DOI matches its Crossref record; 89 of the 89 references it cites were found in a registry or an index; 29 of the 60 sources its claims cite were checked for support.
Academic Integrity considerations of AI Large Language Models in the post-pandemic era: ChatGPT and beyond
Open accessAbstract
This paper explores the academic integrity considerations of students’ use of Artificial Intelligence (AI) tools using Large Language Models (LLMs) such as ChatGPT in formal assessments. We examine the evolution of these tools, and highlight the potential ways that LLMs can support in the education of students in digital writing and beyond, including the teaching of writing and composition, the possibilities of co-creation between humans and AI, supporting EFL learners, and improving Automated Writing Evaluations (AWE). We describe and demonstrate the potential that these tools have in creating original, coherent text that can avoid detection by existing technological methods of detection and trained academic staff alike, demonstrating a major academic integrity concern related to the use of these tools by students. Analysing the various issues related to academic integrity that LLMs raise for both Higher Education Institutions (HEIs) and students, we conclude that it is not the student use of any AI tools that defines whether plagiarism or a breach of academic integrity has occurred, but whether any use is made clear by the student. Deciding whether any particular use of LLMs by students can be defined as academic misconduct is determined by the academic integrity policies of any given HEI, which must be updated to consider how these tools will be used in future educational environments.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
29 of 60 cited sources checked · 15 supported · 0 flagged · accuracy not yet measured
Every source the paper cites for this claim was checked and supports it.
In the paper
“there is the potential for LLMs to produce text that lacks semantic coherence (Dale, 2021)”
What the sources say
- Supportedcites Dale, 2021 source
Section 3.3, “From editor to author,” says earlier language models’ longer outputs could degenerate into nonsense and that GPT-3 can still generate nonsense. This supports the claim that LLMs can produce text lacking semantic coherence. (LLM, conf=0.91)
Strength:hedged· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passes- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“as well as increases in student or academic staff perceptions of AD (Amzalag et al., 2021; Reedy et al., 2021; Walsh et al., 2021) occurring.”
What the sources say
- Partly supportedcites Amzalag et al., 2021 sourceExperimental — accuracy not yet measured
The Findings (page unavailable) report students’ and lecturers’ perceptions of academic dishonesty during pandemic-era online exams, including lecturers’ concern about copying. They compare the two groups, not perceptions before and during the pandemic, so they do not establish that the pandemic increased those perceptions. (LLM, conf=0.86)
- Supportedcites Reedy et al., 2021 source
The Findings report that academic staff thought online exams made cheating easier and expressed concern about an increase in student cheating during the pandemic-driven move to digital assessment. This supports the claim about staff perceptions, not an increase in actual dishonesty. (LLM, conf=0.88)
- Supportedcites Walsh et al., 2021 source
The Conclusion reports that surveyed science undergraduates perceived cheating as more frequent online during the first COVID-interrupted semester. It attributes this perception primarily to the shift to online learning, rather than to the pandemic’s added stresses; the Introduction also describes faculty perceptions of increased cheating during that period. (LLM, conf=0.89)
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passes- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“given the misrepresentation that have been made, usage of this type could be seen as a relatively clear case of plagiarism by the student”
What the sources say
- Partly supportedcites Roe & Perkins, 2022 sourceExperimental — accuracy not yet measured
In “Types of APTs and their use in academic work” (page unavailable), the authors call using automated paraphrasing tools to rework copied source material academic dishonesty and paraphrasing plagiarism. That supports the general concern about presenting machine-assisted work as one’s own, but the paper does not address LLM-generated work or establish that directly copying undisclosed LLM output is a clear case of plagiarism. (LLM, conf=0.85)
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“current research suggests that use of the created text by students is unlikely to be spotted by existing text-matching software tools used by HEIs”
What the sources say
- Partly supportedcites Wahle, Ruas, Kirstein, et al., 2022 sourceExperimental — accuracy not yet measured
In Q5, the paper reports that PlagScan could not reliably identify machine-paraphrased text and that large language models made text matches harder to find. Its tests concern LLM-generated paraphrases and one commercial plagiarism-detection product, not all LLM-created text or the range of text-matching tools used by higher education institutions. (LLM, conf=0.86)
Load-bearingStrength:hedged· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- Rests on an assumption: LLM-generated text is uniquely created based on the inputs provided.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“the use of LLMs in these tools can particularly reduce the burden on EFL instructors (Godwin-Jones, 2022).”
What the sources say
- Partly supportedcites Godwin-Jones, 2022 sourceExperimental — accuracy not yet measured
The Ecological Perspectives section (page unavailable) says AWE can free teachers’ time from correcting language mechanics so they can address higher-order writing issues. The Introduction describes large language models as enabling language services, but the paper does not specifically say that LLMs in AWE tools reduce instructors’ workload, or that the benefit is particular to EFL instructors. (LLM, conf=0.84)
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passes- Partly supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“emerging research has suggested that existing LLMs can produce output which humans struggle to identify as being machine created”
What the sources say
- Not checked — no full text of the sourcecites Abd-Elaal et al., 2022 source
- Supportedcites Clark et al., 2021 source
Section 2.5 reports that evaluators had difficulty identifying GPT-3-generated text. The paper also states that evaluators often mistook machine-generated text for human-authored text, supporting the claim that humans can struggle to identify such output. (LLM, conf=0.93)
- Not checked — no full text of the sourcecites Gunser et al., 2021 source
Load-bearingStrength:hedged· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesThe cited sources support only part of this claim — read each source’s verdict below.
In the paper
“it is highly likely that LLMs will be integrated more deeply into suites of products that are often provided to students by HEIs in the future”
What the sources say
- Partly supportedcites Dale & Viethen, 2021 sourceExperimental — accuracy not yet measured
In the concluding “Copy Shark” passage (page unavailable), the paper predicts that GPT-3-based writing capabilities may soon appear in Microsoft Word. That supports future LLM integration into a widely used product, but the paper does not say that such integration is highly likely across product suites provided to students by higher education institutions. (LLM, conf=0.86)
Load-bearingStrength:hedged· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Gayed et al., (2022) demonstrate how an LLM tool based on GPT-2 can be helpful for EFL students writing English by suggesting predictions which expands on the more basic support offered by existing writing tools.”
What the sources say
- Supportedcites Gayed et al., (2022) source
The Introduction (page unavailable) identifies AI KAKU as an EFL writing aid with a GPT-2-based word-suggestion engine, alongside discussion of traditional spell- and grammar-check tools. The Conclusion (page unavailable) says its descriptive results point to potential usefulness, while noting that the study did not establish a significant difference from the control condition. That supports the claim's cautious wording, “can be helpful,” rather than a claim of proven improvement. (LLM, conf=0.89)
Strength:boosted· not yet calibratedCausal:conditional· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passes- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Their results showed that evaluators were able to identify GPT-2 produced text at an accuracy rate of 57.9%, but GPT-3 at a rate of only 49.9%.”
What the sources say
- Supportedcites Clark et al. (2021) source
Section 2.4 (Results) reports that evaluators correctly distinguished human from GPT-2 text 57.9% of the time, versus 49.9% for human versus GPT-3 text. These are the overall classification accuracies reported in the claim. (LLM, conf=0.94)
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passes- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“as well as increases in student or academic staff perceptions of academic dishonesty occurring (Amzalag et al., 2021; Reedy et al., 2021; Walsh et al., 2021).”
What the sources say
- Partly supportedcites Amzalag et al., 2021 sourceExperimental — accuracy not yet measured
The Findings (page unavailable) report lecturers’ concern about copying and examine students’ and lecturers’ perceptions of dishonesty during pandemic-era online exams. The paper does not report that either group’s perception of dishonesty increased relative to before the pandemic, so the claimed increase is unsupported. (LLM, conf=0.82)
- Supportedcites Reedy et al., 2021 source
The Findings report that staff at the three Australian universities were concerned about an increase in student cheating with the pandemic-driven move from invigilated to online exams. This supports the claim about academic staff perceptions, rather than establishing that cheating itself increased. (LLM, conf=0.87)
- Supportedcites Walsh et al., 2021 source
The Introduction (Student perceptions) says students grew more concerned that peers were cheating after classes moved online due to the pandemic. The Conclusion likewise reports that most surveyed students believed cheating occurred more frequently online, while clarifying that the shift in teaching modality—not the pandemic’s added stresses—was the main reason for that perception. (LLM, conf=0.87)
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passes- Partly supported
- Probably not a substantial claim. 47
- It may state two separate claims as one. 18
Every source the paper cites for this claim was checked and supports it.
In the paper
“These tools started to achieve prominence from 2017 with the emergence of a range of transformer based machine learning models which allow for improved performance on language based tasks when compared to previous machine learning techniques (Vaswani et al., 2017).”
What the sources say
- Supportedcites Vaswani et al., 2017 source
The Abstract (p. 1) introduces the Transformer and reports superior results on two machine-translation tasks. Section 6.1 (p. 8) compares its translation performance with previously published models, supporting the claim that transformer-based models can improve performance on language tasks. (LLM, conf=0.88)
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Digital Writing Assistants (DWAs) such as Grammarly, WordTune and Perusall, use Artificial Intelligence (AI) to improve student writing (Fitria, 2021)”
What the sources say
- Partly supportedcites Fitria, 2021 sourceExperimental — accuracy not yet measured
The Abstract (page not specified) describes Grammarly as an AI-powered writing assistant and reports improved writing performance after its use. The paper supports that part of the claim, but the passages found for WordTune and Perusall do not address either tool, so it does not substantiate the claim about all three. (LLM, conf=0.88)
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“These tools have also demonstrated a particular ability to support English as a Foreign Language (EFL) writers in improving their skills at expressing ideas in written English (Gayed et al., 2022; Nazari et al., 2021; Zhao, 2022)”
What the sources say
- Partly supportedcites Gayed et al., 2022 sourceExperimental — accuracy not yet measured
The Results section reports that adult EFL participants using AI KAKU produced sentences with significantly greater syntactic complexity, which the authors link to communicating complex ideas more fluently. But the study tested performance during a writing task, not lasting improvement in writing skills or a particular ability of digital writing aids generally. Differences in word production and lexical diversity were not statistically significant. (LLM, conf=0.86)
- Partly supportedcites Nazari et al., 2021 sourceExperimental — accuracy not yet measured
The Discussion says Grammarly can improve EFL writing skills, and the Conclusion describes it as an assistive tool for non-native postgraduate students writing in English. But the trial reports improvements in writing self-efficacy, engagement, and emotions—not a measured improvement in students’ ability to express ideas in writing. The Introduction’s reference to expressing ideas describes writing skills generally, not a demonstrated effect of the tool. (LLM, conf=0.84)
- Supportedcites Zhao, 2022 source
The Abstract (p. 2) says Wordtune can help EFL writers learn useful ways to express their ideas in written English. The Conclusion (p. 6) also says it can help them formulate ideas in English and improve writing quality, supporting the attributed point. (LLM, conf=0.88)
Strength:boosted· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Although one concern levied against LLMs is that they will generally always provide an answer or output, even if that is factually incorrect or misinterpreted (Dale, 2021; Hutson, 2022)”
What the sources say
- Partly supportedcites Dale, 2021 sourceExperimental — accuracy not yet measured
Section 3.3, “From editor to author” (page unavailable), says GPT-3 can generate convincing text that is sometimes nonsense, supporting the concern that an output can be unreliable. It does not say LLMs generally always provide an answer, or specifically discuss factual errors or misinterpretation. (LLM, conf=0.82)
- Not checked — no full text of the sourcecites Hutson, 2022 source
Strength:plain· not yet calibratedTests✗ Asserted: fails✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The authors may not put it forward as true. The paper’s conclusions may not depend on it.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Their study suggested that this tool could support in improving the lexical diversity of student work, but statistically significant results were not obtained.”
What the sources say
- Supportedcites Gayed et al., (2022) source
The Results section reports higher average lexical diversity under the AI KAKU condition, which the authors say may indicate improved performance. It also reports that neither lexical-diversity measure showed a statistically significant difference. The Treatment tool section identifies AI KAKU as using GPT-2. (LLM, conf=0.94)
Strength:hedged· not yet calibratedCausal:conditional· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“There are legitimate uses of these tools in supporting the education of students, meaning HEIs must carefully consider how policies dealing with student use of this software are created.”
What the sources say
- Not checked — no full text of the sourcecites Chen et al., 2015 source
- Not checked — no full text of the sourcecites Park & Yang, 2020 source
- Partly supportedcites Zhao, 2022 sourceExperimental — accuracy not yet measured
The Affordance for English Writing section (p. 4) says Wordtune offers learning opportunities for EFL writers and suggests that teachers encourage students to use it in writing activities. This supports an educational use of an AI tool, but our copy does not identify Wordtune as using an LLM. The sentence’s policy recommendation is the citing authors’ inference, not a point attributed to this source. (LLM, conf=0.79)
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✗ Checkable: fails✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may be too vague to check against evidence.
Every source the paper cites for this claim was checked and supports it.
In the paper
“highlight the ethical challenges of deploying any such detectors which may potentially discriminate against EFL students by incorrectly identifying human created text as machine written”
What the sources say
- Supportedcites Fröhling & Zubiaga (2021) source
The Discussion and future work section (page unavailable) warns that detectors could discriminate against non-native speaking groups if they fail to recognize their writing as human. The detection problem section (page unavailable) defines false positives as human-written text incorrectly labelled machine-generated. The paper raises this as a potential risk, not a measured outcome among EFL students. (LLM, conf=0.89)
Strength:hedged· not yet calibratedCausal:conditional· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“However, this tool was tested against GPT-2 produced output and used students rather than academic staff as their participants.”
What the sources say
- Supportedcites Gehrmann et al. (2019) source
The Human-Subjects Study section says the participants were 35 volunteer students in a college-level NLP class, and that each participant assessed texts generated by GPT-2. This supports both parts of the claim; the section also notes that the study included texts from other sources. (LLM, conf=0.94)
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“Solaiman et al. (2019) and Ippolito et al.(2020) also both present tools which showed an encouraging ability in detecting machine created text, although these tools were not tested against the latest generation of LLMs.”
What the sources say
- Supportedcites Solaiman et al. (2019) source
In “Our Work: 1.5 Billion Parameter Model Detection Research” (page not provided), Solaiman et al. describe releasing a detector and report approximately 95% accuracy on GPT-2-generated text. The reported tests use GPT-2 rather than later generations of LLMs. This verdict concerns Solaiman et al. only, not the separately cited Ippolito et al. work. (LLM, conf=0.90)
- Supportedcites Ippolito et al.(2020) source
The Automatic Detection Method section describes the paper’s BERT classifier for distinguishing human- from machine-generated text. Automatic Detection Results reports accuracy of about 81% on the longest excerpts for some sampling methods, while Dataset Methodology says all generated samples came from GPT-2. This supports the detection result and the qualification that the reported tests did not cover newer LLMs. (LLM, conf=0.90)
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
- Supported
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Although in text citations and references were provided (including references from the input text), these were all fabricated, with titles, URLs, and journal names being incorrect.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
None of the sources cited for this claim have been checked yet.
In the paper
“Again, the references were falsified and there were some inaccuracies related to the claimed ability of Turnitin to detect AI generated content, but this fluent, five-paragraph output is typical of a response by ChatGPT.”
Strength:plain· not yet calibratedSingle caseTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“A major benefit of ChatGPT as opposed to previous versions of the GPT-3 software is that of the iteration possible through natural language 'discussions' held with the software.”
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“The output retained the language style used in the input and expanded on the discussions by providing examples specifically related to education as requested.”
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“These tools use Large Language Models (LLMs) and can be used to create original written content that students may use in their assessments.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“ChatGPT in particular can also be challenged if the user believes there are incorrect elements in the output, and can adjust its response accordingly based on any corrections provided by the user.”
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“This output was created in approximately 5 seconds at zero cost (ChatGPT is currently in a free research preview stage and costs are likely to change in the near future).”
Strength:plain· not yet calibratedSingle caseTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“but this fluent, five-paragraph output is typical of a response by ChatGPT.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✗ Checkable: fails✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Deciding whether any particular use of LLMs by students can be defined as academic misconduct is determined by the academic integrity policies of any given HEI, which must be updated to consider how these tools will be used in future educational environments.”
- Also stated in: § Practitioner Notes · p. 2 · hedged
“however it may be a breach of academic integrity policies of any given Higher Education Institution (HEI).”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. The paper may give no reason why it holds.
- Also stated in: § Practitioner Notes · p. 2 · hedged
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“we conclude that it is not the student use of any AI tools that defines whether plagiarism or a breach of academic integrity has occurred, but whether any use is made clear by the student.”
- Also stated in: § Practitioner Notes · p. 2 · hedged
“The use of these tools may not necessarily be considered as plagiarism if students are transparent in how they have been used in any submission, however it may be a breach of academic integrity policies of any given Higher Education Institution (HEI).”
Tests✓ Asserted: passes✗ Checkable: fails✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
- Also stated in: § Practitioner Notes · p. 2 · hedged
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“we conclude that the use of LLMs or LLM based tools should not be considered as plagiarism or a breach of academic integrity if how these tools have been used is stated clearly and made transparent by students.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✗ Checkable: fails✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may be too vague to check against evidence.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This does open a potential avenue of detection by academic staff if there is a suspicion that these tools may have been used by students.”
Strength:hedged· not yet calibratedCausal:conditional· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“These tools can be accessed using commercial services built on this software, often targeted to students as a means of 'assisting' students with assessments.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Many of these tools integrate features of DWAs and APTs, and market the product package together as one 'writing support tool' or similar, available as both free and paid for versions.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“it is feasible to believe that a student could very easily and inadvertently breach any policies related to their 'correct' usage, especially when LLMs become even more deeply integrated into DWAs.”
Strength:hedged· not yet calibratedCausal:conditional· not yet calibratedTests✓ Asserted: passes✗ Checkable: fails✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
- Rests on an assumption: Many advanced DWAs have multiple features available to students which already claim to integrate AI technologies.
- Rests on an assumption: Write Full is a full-service DWA which has integrated paraphrasing tools that may encourage students to use the work of others as inputs, rather than their own writing.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Student usage of an LLM or LLM based tool to create an entire essay, report, or other assignment without stating where this has come from, or one that is presented in a way that was trying to mislead the reader through false citations, would be even more clearly a case of plagiarism, and likely also a broader breach of academic integrity principles.”
Tests✓ Asserted: passes✗ Checkable: fails✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may be too vague to check against evidence. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Tools using Artificial Intelligence (AI) to support in the creation of original text have emerged in recent years, in the form of Large Language Models (LLMs) which are able to develop significant amounts of brand-new text based on short input requests.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, given that the authors and dates provided in the intext citations are correct, this may go unnoticed unless thoroughly checked.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“If the work carried out by a student hasn't been misrepresented, then it does not fall under the strict definition of plagiarism, but it may be considered as academic misconduct depending on the specific policies and regulations in place at individual HEIs.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“A subset of these tools is marketed specifically to students to support in the development of full length-assessments, which also include options to generate citations alongside the work.”
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds. It may state two separate claims as one.
In the paper
“as well as achieving more than 1 million users within five days of launch (Brockman, 2022).”
What the sources say
- Not matchedcites Brockman, 2022
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“HEIs must be aware of the potential danger of learning outcomes not being met if LLMs are being used to develop content, rather than this being created by students themselves.”
Tests✓ Asserted: passes✗ Checkable: fails✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
In the paper
“Dall·E 2 AI system which is able to create and edit detailed digital artwork based on natural language prompts provided by users”
What the sources say
- Not matchedcites OpenAi, n.d.
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“It is evident that with some manipulation of the text through iteration using the tool, along with some minor editing or adjustments of references, a student intending to create large blocks of text to include in an assessment could rapidly complete certain tasks with limited effort and costs.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim.
- Rests on an assumption: The ChatGPT output generated in the paper's demonstration is representative of LLM output generally.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“an alternative method that students may use to access the LLMs is through commercial entities who use an Application Programming Interface (API) to connect to the LLMs. These services are targeted at different groups of customers including copywriters, marketers and students and allow user-friendly access to the LLMs.”
Strength:hedged· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
The paper states this without a source or data behind it here, so there is nothing to check it against. The quote shows it is in the paper.
In the paper
“Students now have easy access to advanced Artificial Intelligence based tools such as ChatGPT.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. The paper may give no reason why it holds.
None of the sources cited for this claim could be matched to a paper in Crossref, so there is nothing to check it against.
In the paper
“Some of the recognised shortcomings of present AWE tools are the generic nature of the feedback provided (Fu et al., 2022) and an inability to focus feedback on how well a task has been addressed (Allen et al., 2016).”
What the sources say
- Not matchedcites Fu et al., 2022
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not matchedcites Allen et al., 2016
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“LLMs therefore represent a clear potential threat to academic integrity as academic staff may be unable to correctly identify the amount of content produced by a student, and therefore provide an accurate evaluation of a student's comprehension and interpretation of the topic at hand.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
None of the sources cited for this claim have been checked yet.
In the paper
“Using software tools to support in AWE of written text and providing feedback to students has been demonstrated to be of benefit to writing instructors and students alike”
What the sources say
Strength:boosted· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“The broad release to the public in November 2022 of ChatGPT by OpenAI marked a significant increase in the ability of the underlying software to create new text using further refined models (GPT-3.5), coupled with a markedly improved user interface (OpenAI, 2022).”
What the sources say
- Not matchedcites OpenAI, 2022
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it. It may state two separate claims as one.
- Not matched
In the paper
“Using LLMs in the teaching of digital writing therefore becomes a method to support writers in engaging more deeply with a topic, rather than at a surface level.”
What the sources say
- Not checked — no full text of the sourcecites McKnight (2021) source
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“During the period of the pandemic, research has shown that students sought out and adapted to new technologies (Vargo et al., 2021)”
What the sources say
- Not checked — no full text of the sourcecites Vargo et al., 2021 source
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
In the paper
“Although it is unlikely that AI will ever take over the traditional role of a teacher (Cope et al., 2021)”
What the sources say
- Not checked — no full text of the sourcecites Cope et al., 2021 source
Strength:hedged· not yet calibratedTests✓ Asserted: passes✗ Checkable: fails✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
In the paper
“Although it is encouraging to note that training of academic staff resulted in an increased ability of academic staff to correctly identify the text sections, the samples were generated using GPT-2 as opposed to GPT-3.”
What the sources say
- Not checked — no full text of the sourcecites Abd-Elaal et al., (2022) source
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“the likelihood of accurately detecting any usage of these tools by students in their submissions typical academic will likely not improve and may even decrease further as new LLMs are developed.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✓ Based: passesIt may state two separate claims as one.
- Rests on an assumption: New large language models will continue to be developed.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“it is not the use of the tools themselves that defines whether plagiarism or a breach of academic integrity has occurred, but whether any such use is made clear.”
Tests✓ Asserted: passes✗ Checkable: fails✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
In the paper
“What we as academic staff define as plagiarism is rapidly changing due to the social construction of this term (Anson, 2022).”
What the sources say
- Not matchedcites Anson, 2022
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Strength:plain· not yet calibratedCausal:direct· not yet calibratedTests✓ Asserted: passes✗ Checkable: fails✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
- Not matched
None of the sources cited for this claim could be matched to a paper in Crossref, so there is nothing to check it against.
In the paper
“These tools include GPTZero (https://gptzero.me/) and Crossplag AI detect (https://crossplag.com/ai-content-detector/) and do show promise in being able to detect the use of AI generated text.”
What the sources say
- Not matchedcites GPTZero (https://gptzero.me/)
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not matchedcites Crossplag AI detect (https://crossplag.com/ai-content-detector/)
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Strength:hedged· not yet calibratedCausal:conditional· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✗ Based: failsProbably not a substantial claim. The paper’s conclusions may not depend on it. The paper may give no reason why it holds.
- Not matched
In the paper
“Although evidence suggests that plagiarism overall has decreased between 1990 and 2020 (Curtis, 2022)”
What the sources say
- Not matchedcites Curtis, 2022
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Strength:hedged· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. The paper’s conclusions may not depend on it.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“a blanket ban of these tools at an institutional level is neither feasible, nor enforceable.”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✗ Checkable: fails✓ Load-bearing: passes✓ Based: passesProbably not a substantial claim. It may be too vague to check against evidence. It may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Deciding whether any particular use of LLMs by students may be defined as academic misconduct will be determined by the future policies of any given HEI”
Tests✓ Asserted: passes✗ Checkable: fails✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
- Rests on an assumption: HEIs will set future policies governing LLM use.
In the paper
“They found that regardless of background, participants found this to be a challenging task, with a high likelihood of ascribing the AI writing samples to humans.”
What the sources say
- Not matchedcites Kumar et al. (2022)
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
Strength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes? Load-bearing: unclear✓ Based: passesIt may state two separate claims as one.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“the future development of LLMs and broader AI supported digital tools have a strong potential for improving the experiences of students and teachers alike in the next generation of HEI classrooms, both in writing instruction and beyond.”
Strength:boosted· not yet calibratedCausal:conditional· not yet calibratedTests✓ Asserted: passes✗ Checkable: fails✗ Load-bearing: fails✓ Based: passesProbably not a substantial claim. It may be too vague to check against evidence. The paper’s conclusions may not depend on it.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Given that both academic staff, as well as technological methods of detection are unable to accurately detect machine generated text and therefore student uses of LLM based tools”
Load-bearingStrength:plain· not yet calibratedTests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes? Based: unclearIt may state two separate claims as one.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“we believe that the future integration of LLMs and other AI supported digital tools into the classroom environment is highly likely”
Load-bearingStrength:hedged· not yet calibratedTests✓ Asserted: passes✗ Checkable: fails✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. It may be too vague to check against evidence. The paper may give no reason why it holds.
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“the inability to accurately detect whether LLMs have been used presents a clear threat to the academic integrity of HEIs.”
Tests✓ Asserted: passes✓ Checkable: passes✓ Load-bearing: passes✗ Based: failsProbably not a substantial claim. The paper may give no reason why it holds.
The paper relies on this without support.
The ChatGPT output generated in the paper's demonstration is representative of LLM output generally.
“As an example of the fluency of the output that can be created using these tools, the two paragraphs of the section above titled: 'The emergence of Large Language Models' were used as an input to ChatGPT.”
The paper relies on this without support.
LLM-generated text is uniquely created based on the inputs provided.
“given that the text that is produced by LLMs is uniquely created based on the inputs provided”
The paper relies on this without support.
Many advanced DWAs have multiple features available to students which already claim to integrate AI technologies.
“Given that many of the more advanced DWAs have multiple features available to students which already claim to integrate AI technologies”
The paper relies on this without support.
Write Full is a full-service DWA which has integrated paraphrasing tools that may encourage students to use the work of others as inputs, rather than their own writing.
“This already appears to be the case in the example of one DWA called Write Full: a full-service DWA which has integrated paraphrasing tools which may encourage students to use the work of others as inputs, rather than their own writing.”
The paper relies on this without support.
New large language models will continue to be developed.
“as new LLMs are developed.”
The paper relies on this without support.
HEIs will set future policies governing LLM use.
“will be determined by the future policies of any given HEI”
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?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“In response, universities have adopted varied approaches: some have banned generative AI or employed detection tools, while others permit and even encourage its use with appropriate acknowledgment (Perkins 2023; Sullivan, Kelly, and McLaughlan 2023; Xiao, Chen, and Bao 2023).”
From Addressing student non-compliance in AI use declarations: implications for academic integrity and assessment in higher education (2025) · cites this paper as Perkins 2023
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Initially, the use of GenAI tools in academic settings was largely viewed as a form of plagiarism or academic misconduct (Cotton et al., 2023; Perkins, 2023).”
From The Artificial Intelligence Assessment Scale (AIAS): A Framework for Ethical Integration of Generative AI in Educational Assessment (2024) · cites this paper as Perkins, 2023
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Banning or blocking GenAI tools has proven ineffective, and punitive approaches ignore the potential benefits of these technologies.”
From The AI Assessment Scale (AIAS) in action: A pilot implementation of GenAI supported assessment (2024) · cites this paper as Perkins, 2023
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“it leads to a narrow perception of AI as only a source of academic malpractice (Perkins, 2023).”
From Assessment-focused governance mode? Mapping Generative AI policies across 76 UK higher education institutions: Policy framework and recommendations (2026) · cites this paper as Perkins, 2023
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The current study identified critical thinking, self-reflection, and data interpretation as areas where students could work collaboratively with generative AI tools without impinging on institutional academic integrity policies (Perkins, 2023).”
From Assessment design through co-creation (2025) · cites this paper as Perkins, 2023
Experimental — accuracy not yet measured
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“The unilateral deployment of AI detection software could breach students' due process rights, especially in ESL situations where language usage could be similar to AI-produced writing [13].”
From Evaluating the Effectiveness and Ethical Implications of AI Detection Tools in Higher Education (2025) · cites this paper as [13]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“While AI detectors offer scalable solutions, they frequently produce false positives and lack transparency, especially for multilingual or non-native English speakers.”
From Evaluating the Effectiveness and Ethical Implications of AI Detection Tools in Higher Education (2025) · cites this paper as [13]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“However, as multiple studies have shown, their reliability and integration into academic workflows remain inconsistent [10,13].”
From Evaluating the Effectiveness and Ethical Implications of AI Detection Tools in Higher Education (2025) · cites this paper as [13]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“As AI-generated content becomes more prevalent, institutions are revising academic integrity policies and assessment practices [13].”
From Evaluating the Effectiveness and Ethical Implications of AI Detection Tools in Higher Education (2025) · cites this paper as [13]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“These discrepancies confuse students and faculty, leading to uneven enforcement and potential legal risks for institutions that fail to maintain fair and consistent procedures [13].”
From Evaluating the Effectiveness and Ethical Implications of AI Detection Tools in Higher Education (2025) · cites this paper as [13]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Because large language models (LLMs) excel at synthesizing domainspecific technical concepts and mimicking natural linguistic fluency, traditional written tasksuch as take-home assignments, reflective reports, or online language modules-are increasingly vulnerable to unearned authorship and cognitive shortcut (Cotton et al., 2023;Perkins, 2023).”
From Navigating Academic Integrity and Student Rationalisations of Generative AI in Vocational Online Learning Environments (2026) · cites this paper as Perkins, 2023
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Despite advancements in AI detection tools, the effectiveness of conventional anti-plagiarism software, such as iThenticate and Turnitin, remains questionable against AI-generated content (Perkins, 2023;Chaka, 2023Chaka, , 2024;;Hassoulas et al., 2023;Ifelebuegu, 2023;Mohammadkarimi et al., 2023;Sullivan et al., 2023).”
From Higher Education's Generative Artificial Intelligence Paradox: The Meaning of Chatbot Mania (2024) · cites this paper as Perkins, 2023
The cited source says what the claim attributes to it.
In the citing paper
“These updates aim to clarify what constitutes academic misconduct when AI is involved, drawing a line between legitimate support and dishonest automation [13].”
From Evaluating the Effectiveness and Ethical Implications of AI Detection Tools in Higher Education (2025) · cites this paper as [13]
The cited source says what the claim attributes to it.
In the citing paper
“As Cotton, Cotton, & Shipway (2024) note, AI-generated content challenges traditional definitions of authorship and originality, a concern echoed by Perkins (2023).”
From Generative AI-Enhanced Oral Exams in Economics Education (2026) · cites this paper as Perkins (2023)
In the citing paper
“At the same time, experts in our study were aware of the limitations of LLMs and cited similar apprehensions to those raised in the literature (Dwivedi et al. 2023;OpenAI 2022;Perkins 2023).”
From Friend or foe? Exploring the implications of large language models on the science system (2025) · cites this paper as Perkins 2023
In the citing paper
“To date, scholars have primarily focused on the implications of LLMs on education (e.g., Perkins 2023; Fyfe 2023) with limited attention being paid to their impact on science and scientific practices (for exceptions, see Chubb et al. 2022;Ribeiro et al. 2023).”
From Friend or foe? Exploring the implications of large language models on the science system (2025) · cites this paper as Perkins 2023
In the citing paper
“This adds a new and fresh direction to the LLM discussion, which has focused heavily on the impact of this technology for individual academics and learners (Fyfe 2023;Perkins 2023).”
From Friend or foe? Exploring the implications of large language models on the science system (2025) · cites this paper as Perkins 2023
In the citing paper
“chatbots present a substantial challenge concerning maintaining academic integrity [20, 56] and copyright violations [83], which are significant concerns in education”
From A systematic literature review of empirical research on ChatGPT in education (2024) · cites this paper as [56]
In the citing paper
“Indicating that LLMs provide an obvious potential danger to academic integrity since instructors may be unable to accurately determine the quantity of information generated by a student and hence offer an appropriate evaluation of a student’s grasp and interpretation of the issue at hand.”
From ChatGPT Unveiled: Understanding Perceptions of Academic Integrity in Higher Education - A Qualitative Approach (2025) · cites this paper as Perkins (2023)
In the citing paper
“Misusing tools such as ChatGTP for completing non-invigilated assessment tasks and presenting machine-generated outputs as a student's own original work pose a considerable threat to academic honesty, challenging academics and universities to react and to define appropriate guidelines and policies (Perkins, 2023).”
From The rapid rise of generative AI and its implications for academic integrity: Students’ perceptions and use of chatbots for assistance with assessments (2024) · cites this paper as Perkins, 2023
Citations
Cites
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404
Jakubův, J., Chvalovský, K., Goertzel, Z., et al. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.itp.2023.19 · Cited by 76,648
On the Opportunities and Risks of Foundation Models
Bommasani, R., Hudson, D. A., Adeli, E., Altman, R., Arora, S., von Arx, S., Bernstein, M. S., Bohg, J., Bosselut, A., Brunskill, E., et al. (2021) · arXiv · doi:10.48550/arxiv.2108.07258 · Cited by 2,676
Risko, E. F., & Gilbert, S. J. (2016) · Trends in Cognitive Sciences · doi:10.1016/j.tics.2016.07.002 · Cited by 1,502 · cited 1×
“Therefore, we may consider LLMs as another potential tool for reducing the cognitive demands required by a task, described by Risko and Gilbert (2016) as a process called cognitive offloading.”
Balancing Technology, Pedagogy and the New Normal: Post-pandemic Challenges for Higher Education
Rapanta, C., Botturi, L., Goodyear, P., Ortiz, L. G., & Koole, M. L. (2021) · Postdigital Science and Education · doi:10.1007/s42438-021-00249-1 · Cited by 726
Multitask Prompted Training Enables Zero-Shot Task Generalization
Sanh, V., Webson, A., Raffel, C., Bach, S. H., Sutawika, L., Alyafeai, Z., Chaffin, A., Stiegler, A., Scao, T. L., Raja, A., et al. (2021) · arXiv · doi:10.48550/arxiv.2110.08207 · Cited by 612
Artificial intelligence for education: Knowledge and its assessment in AI-enabled learning ecologies
Cope, B., Kalantzis, M., & Searsmith, D. (2021) · Educational Philosophy and Theory · doi:10.1080/00131857.2020.1728732 · Cited by 518
Dale, R. (2020) · Natural Language Engineering · doi:10.1017/s1351324920000601 · Cited by 509
GLTR: Statistical Detection and Visualization of Generated Text
Gehrmann, S., Strobelt, H., & Rush, A. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics: System Demonstrations · doi:10.18653/v1/p19-3019 · Cited by 470 · cited 2×
“Technological methods to identify text generated by LLMs have been proposed by Gehrmann et al. (2019) who present the results of a study using a tool named GLTR.”
GLTR: Statistical Detection and Visualization of Generated Text
Gehrmann, S., Strobelt, H., & Rush, A. M. (2019) · arXiv · doi:10.48550/arxiv.1906.04043 · Cited by 460
Digital technology use during COVID ‐19 pandemic: A rapid review
Vargo, D., Zhu, L., Benwell, B., & Yan, Z. (2021) · Human Behavior and Emerging Technologies · doi:10.1002/hbe2.242 · Cited by 459
Köbis, N., & Mossink, L. D. (2021) · Computers in Human Behavior · doi:10.1016/j.chb.2020.106553 · Cited by 406
Nazari, N., Shabbir, M. S., & Setiawan, R. (2021) · Heliyon · doi:10.1016/j.heliyon.2021.e07014 · Cited by 405
A taxonomy of prompt modifiers for text-to-image generation
Oppenlaender, J. (2023) · Behaviour and Information Technology · doi:10.1080/0144929x.2023.2286532 · Cited by 385
Lee, M., Liang, P., & Yang, Q. (2022) · CHI Conference on Human Factors in Computing Systems · doi:10.1145/3491102.3502030 · Cited by 345 · cited 1×
“…AI based technologies, there is ongoing research exploring how the intricacies of co-creation may be further be codified and explored in the fields of creative writing (M. Lee et al., 2022; Y. Lee et al., 2022), software development (Biderman & Raff, 2022) and artistic outputs (Oppenlaender, 2022).”
Exploring an AI-based writing Assistant's impact on English language learners
Gayed, J. M., Carlon, M. K. J., Oriola, A. M., & Cross, J. S. (2022) · Computers and Education: Artificial Intelligence · doi:10.1016/j.caeai.2022.100055 · Cited by 340
Contract cheating by STEM students through a file sharing website: a Covid-19 pandemic perspective
Lancaster, T., & Cotarlan, C. (2021) · International Journal for Educational Integrity · doi:10.1007/s40979-021-00070-0 · Cited by 312 · cited 2×
“…the COVID-19 pandemic has caused an increase in both the detected number of academic dishonesty cases (Henderson et al., 2022; Jenkins et al., 2022; Lancaster & Cotarlan, 2021), as well as increases in student or academic staff perceptions of academic dishonesty occurring (Amzalag et al., 2021;…”
Digital support for academic writing: A review of technologies and pedagogies
Strobl, C., Ailhaud, E., Benetos, K., Devitt, A., Kruse, O., Proske, A., & Rapp, C. (2019) · Computers & Education · doi:10.1016/j.compedu.2018.12.005 · Cited by 298
Plagiarism in the age of massive generative pre-trained transformers (GPT-3)
Dehouche, N. (2021) · Ethics in Science and Environmental Politics · doi:10.3354/esep00195 · Cited by 296
Release Strategies and the Social Impacts of Language Models
Solaiman, I., Brundage, M., Clark, J., Askell, A., Herbert-Voss, A., Wu, J., Radford, A., Krueger, G., Kim, J. W., Kreps, S., et al. (2019) · arXiv · doi:10.48550/arxiv.1908.09203 · Cited by 292
All That’s ‘Human’ Is Not Gold: Evaluating Human Evaluation of Generated Text
Clark, E., August, T., Serrano, S., Haduong, N., Gururangan, S., & Smith, N. A. (2021) · Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) · doi:10.18653/v1/2021.acl-long.565 · Cited by 290
All That's 'Human' Is Not Gold: Evaluating Human Evaluation of Generated Text
Clark, E., August, T., Serrano, S., Haduong, N., Gururangan, S., & Smith, N. A. (2021) · arXiv · doi:10.48550/arxiv.2107.00061 · Cited by 290
Partnering with AI: Intelligent writing assistance and instructed language learning
Godwin-Jones, R. (2022) · Language Learning & Technology · doi:10.64152/10125/73474 · Cited by 285
Automatic Detection of Generated Text is Easiest when Humans are Fooled
Ippolito, D., Duckworth, D., Callison-Burch, C., & Eck, D. (2019) · arXiv · doi:10.48550/arxiv.1911.00650 · Cited by 264 · cited 1×
“Solaiman et al. (2019) and Ippolito et al.(2020) also both present tools which showed an encouraging ability in detecting machine created text, although these tools were not tested against the latest generation of LLMs. Fröhling & Zubiaga (2021) present a promising low cost detection model which is…”
Automatic Detection of Generated Text is Easiest when Humans are Fooled
Ippolito, D., Duckworth, D., Callison-Burch, C., & Eck, D. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.164 · Cited by 264
Automated Writing Assessment in the Classroom
Warschauer, M., & Grimes, D. (2008) · Pedagogies: An International Journal · doi:10.1080/15544800701771580 · Cited by 243
Cheating in the Digital Age: Do Students Cheat More in Online Courses?.
Watson, G., & Sottile, J. (2010) · Online journal of distance learning administration · Cited by 242
The great Transformer: Examining the role of large language models in the political economy of AI
Luitse, D. M. R., & Denkena, W. (2021) · Big Data & Society · doi:10.1177/20539517211047734 · Cited by 242
Could AI help you to write your next paper?
Hutson, M. (2022) · Nature · doi:10.1038/d41586-022-03479-w · Cited by 236
Academic Integrity and Distance Learning
Lanier, M. M. (2006) · Journal of Criminal Justice Education · doi:10.1080/10511250600866166 · Cited by 226
Grammarly as AI-powered English Writing Assistant: Students' Alternative for Writing English
Fitria, T. N. (2021) · Metathesis: Journal of English Language, Literature, and Teaching · doi:10.31002/metathesis.v5i1.3519 · Cited by 221
Academic Honesty and Online Courses
Grijalva, T. C., Nowell, C., & Kerkvliet, J. R. (2006) · Cited by 220
A review of AWE feedback: types, learning outcomes, and implications
Fu, Q.-K., Zou, D., Xie, H., & Cheng, G. (2022) · Computer Assisted Language Learning · doi:10.1080/09588221.2022.2033787 · Cited by 210
Zhao, X. (2023) · RELC Journal · doi:10.1177/00336882221094089 · Cited by 192
Stop! Grammar time: University students’ perceptions of the automated feedback program Grammarly
ONeill, R., & Russell, A. (2019) · Australasian Journal of Educational Technology · doi:10.14742/ajet.3795 · Cited by 180
Using Internet based paraphrasing tools: Original work, patchwriting or facilitated plagiarism?
Rogerson, A. M., & McCarthy, G. (2017) · International Journal for Educational Integrity · doi:10.1007/s40979-016-0013-y · Cited by 179
Reedy, A., Pfitzner, D., Rook, L., & Ellis, L. (2021) · International Journal for Educational Integrity · doi:10.1007/s40979-021-00075-9 · Cited by 168
Point, Click, and Cheat: Frequency and Type of Academic Dishonesty in the Virtual Classroom
Stuber, D., Wiseley, P., & Hoggatt, S. (2009) · Online journal of distance learning administration · Cited by 147
Beyond "Gotcha!": Situating Plagiarism in Policy and Pedagogy
Price, M. (2002) · College Composition and Communication · doi:10.2307/1512103 · Cited by 118
Amzalag, M., Shapira, N., & Dolev, N. (2022) · Journal of Academic Ethics · doi:10.1007/s10805-021-09413-5 · Cited by 118
Biermann, O. C., Ma, N. F., & Yoon, D. (2022) · Designing Interactive Systems Conference · doi:10.1145/3532106.3533506 · Cited by 116
Feature-based detection of automated language models: tackling GPT-2, GPT-3 and Grover
Fröhling, L., & Zubiaga, A. (2021) · PeerJ Computer Science · doi:10.7717/peerj-cs.443 · Cited by 106
COVID‐19 as the tipping point for integrating e‐assessment in higher education practices
St‐Onge, C., Ouellet, K., Lakhal, S., Dubé, T. V., & Marceau, M. (2021) · British Journal of Educational Technology · doi:10.1111/bjet.13169 · Cited by 105
Paraphrasing tools, language translation tools and plagiarism: an exploratory study
Prentice, F. M., & Kinden, C. E. (2018) · International Journal for Educational Integrity · doi:10.1007/s40979-018-0036-7 · Cited by 99
When Opportunity Knocks: College Students' Cheating Amid the COVID-19 Pandemic
Jenkins, B. D., Golding, J. M., Le Grand, A. M., Levi, M. M., & Pals, A. M. (2023) · Teaching of Psychology · doi:10.1177/00986283211059067 · Cited by 85
Roe, J., & Perkins, M. (2022) · International Journal for Educational Integrity · doi:10.1007/s40979-022-00109-w · Cited by 83
Self-citationThe automated writing assistance landscape in 2021
Dale, R., & Viethen, J. (2021) · Natural Language Engineering · doi:10.1017/s1351324921000164 · Cited by 81
Re-imagining University Assessment in a Digital World
Bearman, M., Dawson, P., Ajjawi, R., Tai, J., & Boud, D. J. (2020) · The enabling power of assessment · doi:10.1007/978-3-030-41956-1 · Cited by 80
(2016) · Grantee Submission · Cited by 79
Fatemi, G., & Saito, E. (2019) · Journal of Further and Higher Education · doi:10.1080/0309877x.2019.1683521 · Cited by 78
Nobles, S., & Paganucci, L. (2015) · Computers & composition/Computers and composition · doi:10.1016/j.compcom.2015.09.001 · Cited by 76
(2016) · Journal of academic language and learning · Cited by 75
Developing a corpus-based paraphrase tool to improve EFL learners' writing skills
Chen, M.-H., Huang, S.-T., Chang, J. S., & Liou, H.-C. (2015) · Computer Assisted Language Learning · doi:10.1080/09588221.2013.783873 · Cited by 72
Identifying Machine-Paraphrased Plagiarism
Wahle, J. P., Ruas, T., Foltýnek, T., Meuschke, N., & Gipp, B. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-96957-8_34 · Cited by 65
Assisting academics to identify computer generated writing
Abd-Elaal, E.-S., Gamage, S. H. P. W., & Mills, J. E. (2022) · European Journal of Engineering Education · doi:10.1080/03043797.2022.2046709 · Cited by 62
Walsh, L. L., Lichti, D. A., Zambrano-Varghese, C. M., Borgaonkar, A. D., Sodhi, J. S., Moon, S., Wester, E. R., & Callis-Duehl, K. L. (2021) · International Journal for Educational Integrity · doi:10.1007/s40979-021-00089-3 · Cited by 61
Teaching Models to Express Their Uncertainty in Words
Lin, S., Hilton, J., & Evans, O. (2022) · arXiv · doi:10.48550/arxiv.2205.14334 · Cited by 59
Cognitive Offloading and Assessment
Dawson, P. (2020) · The enabling power of assessment · doi:10.1007/978-3-030-41956-1_4 · Cited by 53
Cheaters on Twitter: an analysis of engagement approaches of contract cheating services
Amigud, A. (2019) · Studies in Higher Education · doi:10.1080/03075079.2018.1564258 · Cited by 50
Perkins, M., Gezgin, U. B., & Roe, J. (2018) · Journal of Academic Ethics · doi:10.1007/s10805-018-9311-8 · Cited by 47
Self-citationSun, Y.-C. (2013) · Journal of English for Academic Purposes · doi:10.1016/j.jeap.2013.07.002 · Cited by 45
Kaqinari, T., Makarova, E., Audran, J., Döring, A. K., Göbel, K., & Kern, D. (2021) · Journal of University Teaching and Learning Practice · doi:10.53761/1.18.5.10 · Cited by 44
Prompt Engineering for Text-Based Generative Art
Oppenlaender, J. (2022) · arXiv.org · doi:10.48550/arxiv.2204.13988 · Cited by 44
Factors associated with online examination cheating
Henderson, M., Chung, J., Awdry, R., Mundy, M., Bryant, M., Ashford, C., & Ryan, K. (2023) · Assessment & Evaluation in Higher Education · doi:10.1080/02602938.2022.2144802 · Cited by 43
Fooling MOSS Detection with Pretrained Language Models
Biderman, S., & Raff, E. (2022) · arXiv · doi:10.48550/arxiv.2201.07406 · Cited by 37
Fooling MOSS Detection with Pretrained Language Models
Biderman, S., & Raff, E. (2022) · Proceedings of the 31st ACM International Conference on Information & Knowledge Management · doi:10.1145/3511808.3557079 · Cited by 37
Electric Sheep? Humans, Robots, Artificial Intelligence, and the Future of Writing
McKnight, L. (2021) · Changing English · doi:10.1080/1358684x.2021.1941768 · Cited by 36
How Large Language Models are Transforming Machine-Paraphrase Plagiarism
Wahle, J. P., Ruas, T., Kirstein, F., & Gipp, B. (2022) · Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2022.emnlp-main.62 · Cited by 35
Are Neural Language Models Good Plagiarists? A Benchmark for Neural Paraphrase Detection
Wahle, J. P., Ruas, T., Meuschke, N., & Gipp, B. (2021) · ACM/IEEE Joint Conference on Digital Libraries · doi:10.1109/jcdl52503.2021.00065 · Cited by 35
Computer-based writing tools and the cognitive needs of novice writers
Kozma, R. (1991) · Computers & composition/Computers and composition · doi:10.1016/8755-4615(91)80046-g · Cited by 30
Interactive Children’s Story Rewriting Through Parent-Children Interaction
Lee, Y.-J., Kim, T. S., Chang, M., & Kim, J. (2022) · Workshop on Intelligent and Interactive Writing Assistants (In2Writing) · doi:10.18653/v1/2022.in2writing-1.9 · Cited by 26
Gunser, V. E., Gottschling, S., Brucker, B., Richter, S., & Gerjets, P. (2021) · Communications in Computer and Information Science · doi:10.1007/978-3-030-78635-9_67 · Cited by 25
DSpin: Detecting Automatically Spun Content on the Web
Zhang, Q., Wang, D. Y., & Voelker, G. M. (2014) · Network and Distributed System Security Symposium · doi:10.14722/ndss.2014.23004 · Cited by 24
Neural automated writing evaluation for Korean L2 writing
Lim, K., Song, J., & Park, J. (2022) · Natural Language Engineering · doi:10.1017/s1351324922000298 · Cited by 23
How Large Language Models are Transforming Machine-Paraphrased PlagiarismPreprint
Wahle, J. P., Ruas, T., Kirstein, F., & Gipp, B. (2023) · doi:10.22541/au.167528149.95939897/v1 · Cited by 12
Parkinson, A. L., Hatje, E., Kynn, M., Kuballa, A., Donkin, R., & Reinke, N. B. (2022) · Assessment & Evaluation in Higher Education · doi:10.1080/02602938.2022.2040947 · Cited by 11
Plagiarism in higher education: classification, causes and controls
Perkins, M., Gezgin, U. B., & Gordon, R. D. (2019) · doi:10.13140/rg.2.2.20694.11841 · Cited by 9
Self-citationUtilizing an AI-Based Grammar Checker in an EFL Writing Classroom
Park, J.-H., & Yang, I. Y. (2020) · Korean Journal of Applied Linguistics · doi:10.17154/kjal.2020.3.36.1.97 · Cited by 8
Information for a Better World: Shaping the Global Future
(2022) · Lecture notes in computer science · doi:10.1007/978-3-030-96957-8 · Cited by 7
Artificial Intelligence, Algorithmic Writing & Educational Ethics
(2021) · Cited by 4
HCI International 2021 - Posters
(2021) · Communications in computer and information science · doi:10.1007/978-3-030-78635-9 · Cited by 2
Model Agnostic Interpretability for Multiple Instance Learning
Early, J., Evers, C., & Ramchurn, S. D. (2022) · ePrints Soton (University of Southampton) · doi:10.48550/arxiv.2201.11701 · Cited by 2
OPT: Open Pre-trained Transformer Language Models
Zhang, S., Roller, S., Goyal, N., Artetxe, M., Chen, M., Chen, S., Dewan, C., Diab, M., Li, X., Lin, X. V., et al. (2022) · arXiv · doi:10.48550/arxiv.2205.01068
How Large Language Models are Transforming Machine-Paraphrased Plagiarism
Wahle, J. P., Ruas, T., Kirstein, F., et al. (2022) · arXiv · doi:10.48550/arxiv.2210.03568
Flagged first, then by citations
Research works printed but not linked: 6
- Curtis, G. J. (2022). Trends in plagiarism and cheating prevalence: 1990-2020 and beyond. In D. Rettinger & T. Bertram Gallant (Eds.), Cheating academic integrity (pp. 11-44). Jossey-Bass. https://www.wiley.com/enau/Cheating+Academic+Integrity%3A+Lessons+from+30+Years+of+Research-p9781119868170
- Dinneen, C. (2021). Students' use of digital translation and paraphrasing tools in written assignments on Direct Entry English Programs. English Australia Journal , 37 (1), 40-51. https://doi.org/10.3316/informit.748262877947586 · DOI
- Kumar, R., Mindzak, M., Eaton, S. E., & Morrison, R. (2022). AI & AI: Exploring the contemporary intersections of artificial intelligence and academic integrity . Canadian Society for the Study of Higher Education Annual Conference, Online. Werklund School of Education. https://tinyurl.com/ycknz8fd
- Kumar, R., Mindzak, M., & Racz, R. (2022). Who Wrote This? The Use of Artificial Intelligence in the Academy . http://hdl.handle.net/10464/16532
- Miller, A., & Young-Jones, A. D. (2012). Academic integrity: Online classes compared to face-toface classes. Journal of Instructional Psychology , 39 (3-4), Article 3-4. https://tinyurl.com/mrxnp9zy
- Wilder, N., Weßels, D., Gröpler, J., Klein, A., & Mundorf, M. (2021). Who is Responsible for Integrity in the Age of Artificial Intelligence? An Analysis Using the Example of Academic Writing. European Conference on Academic Integrity and Plagiarism 2021 , 179-181. https://tinyurl.com/2s4r5dyr
Cited by
Sorted by citations
What Is the Impact of ChatGPT on Education? A Rapid Review of the Literature
Lo, C. K. (2023) · Education Sciences · doi:10.3390/educsci13040410 · Cited by 2,024
New Era of Artificial Intelligence in Education: Towards a Sustainable Multifaceted Revolution
Kamalov, F., Santandreu Calonge, D., & Gurrib, I. (2023) · Sustainability · doi:10.3390/su151612451 · Cited by 1,089
“The new technology is “a clear threat to academic integrity for HEIs, requiring a range of adjustments to be made in both practice and policy” [53].”
A survey on large language model (LLM) security and privacy: The Good, The Bad, and The Ugly
Yao, Y., Duan, J., Xu, K., Cai, Y., Sun, Z., & Zhang, Y. (2024) · High-Confidence Computing · doi:10.1016/j.hcc.2024.100211 · Cited by 1,083
Grassini, S. (2023) · Education Sciences · doi:10.3390/educsci13070692 · Cited by 1,048
Strzelecki, A. (2024) · Interactive Learning Environments · doi:10.1080/10494820.2023.2209881 · Cited by 807
Metadata
- Authors
- Perkins, M.
- Affiliations
- British University Vietnam (VN)
- Corresponding author
- Perkins, Mike
- ORCID
- Perkins, Mike
- Venue
- Journal of University Teaching and Learning Practice
- Publisher
- Open Access Publishing Association
- Published
- 2023-01-01
- Volume
- 20
- Issue
- 2
- ISSN
- 1449-9789 (ISSN-L)
- OpenAlex
- W4321605350
- Cited by
- 839
- References
- 84
- Citations by year
- 2026: 270 · 2025: 210 · 2024: 251 · 2023: 108
- FWCI
- 209.33
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Locations
- Journal of University Teaching and Learning Practice (publishedVersion)
- Research Online (University of Wollongong) (submittedVersion)
- Topics
- Social Sciences › Education › Academic integrity and plagiarism
- Social Sciences › Education › Artificial Intelligence in Education
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Keywords
- academic integrity, large language models, ChatGPT, generative AI in education, academic misconduct, plagiarism, automated writing evaluation, EFL learners, digital writing, composition teaching, human-AI co-creation, academic integrity policy, higher education, post-pandemic education
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
