Skip to content

WHITEPAPER

Beyond Detection

A Framework for Ethical AI Integration in Academic Research

Alex SpezowkaJenni Research2026References (81)

  • <80%AI-detector accuracy in independent testing
  • 50%False positives on widely used platforms
  • 61%Of non-native English writing flagged as AI

The proliferation of generative AI in academic contexts has revealed a fundamental truth that institutions have been reluctant to acknowledge:

The detection paradigm has failed.

AI detection tools achieve accuracy rates often below 80% in independent testing.1 Wakjira et al. (2025) False positive rates can be as high as 50% across widely-used platforms.2 Weber-Wulff et al. (2023) There is also documented systematic bias, with over 61% of non-native English writing flagged as AI-generated.3 Liang et al. (2023) The current approach of “detect and punish” thus creates more harm than it prevents.

Meanwhile, the integration of AI into scholarly work has already occurred at scale. Studies indicate that 13.5% to 22.5% of academic papers now show evidence of AI assistance.4 Liang et al. (2025a) Peer review processes are similarly affected, with at least 15.8% of reviews showing AI assistance in some conferences.5 Latona et al. (2024) This whitepaper argues that the question facing higher education is no longer how to prevent AI use, but rather how to build institutional frameworks that make AI integration transparent, bounded, and pedagogically sound.

The path forward requires abandoning unreliable surveillance in favor of transparency architectures—tools and policies designed from inception to make AI contributions visible, auditable, and appropriately constrained. This shift represents not a capitulation to technological inevitability, but an opportunity to re-imagine academic integrity for a new era.

Part I: The Epistemological Limits of AI Detection

The widespread deployment of AI detection tools in academic settings rests on a flawed premise: that AI-generated and human-generated text represent distinguishable categories that can be reliably separated through algorithmic analysis. Empirical evidence accumulated since late 2022 demonstrates that this assumption fails under rigorous testing.6 Weber-Wulff et al. (2023); Sadasivan et al. (2023)

The most comprehensive independent evaluation of AI detection tools to date, conducted by Weber-Wulff and colleagues and published in the International Journal for Educational Integrity, tested fourteen detection tools, including twelve publicly available platforms and two commercial systems.7 Weber-Wulff et al. (2023) The findings were unequivocal: all tools scored below 80% accuracy, with only five exceeding 70%. The study concluded that these tools are neither accurate nor reliable, exhibiting a main bias toward classifying output as human-written rather than detecting AI-generated text.

The performance degradation becomes more severe under realistic conditions. Studies indicate that even with relatively straightforward AI-generated text, detection tools exhibit notable issues, and their accuracy drops significantly when faced with manipulated content or obfuscation techniques such as paraphrasing.8 Perkins et al. (2024c); Weber-Wulff et al. (2023) Some AI content detectors have been found ineffective in identifying paraphrased texts.9 Liu et al. (2024)

Turnitin, a widely used commercial system in academia, has attempted to address limitations through deliberate calibration, reportedly accepting a certain percentage of false negatives to maintain a low false positive rate at the document level.10 Jabarian and Imas (2025) This strategic trade-off, while a product design choice, underscores a fundamental challenge: detection systems often cannot simultaneously achieve high sensitivity (catching AI-generated content) and high specificity (avoiding misidentifying human-written content). The difficulty for traditional AI detectors to reliably differentiate between legitimate human paraphrasing and AI generation undermines their utility for academic integrity enforcement.11 Bordalejo et al. (2025)

The theoretical foundations for this limitation were established by Sadasivan and colleagues at the University of Maryland.12 Sadasivan et al. (2023) Their mathematical analysis demonstrated that for sufficiently capable language models, even optimal detectors can only marginally outperform random classification, especially after applying transformations like paraphrasing.13 Cheng et al. (2025); Sadasivan et al. (2023) As Professor Soheil Feizi summarized, “There is an upper bound on our detectors that fundamentally limits them, so it’s very unlikely that we’ll be able to develop detectors that will reliably identify AI-generated content.”14 University of Maryland (2025)

This is not merely a technical limitation awaiting an engineering solution. As AI language models improve and their outputs converge with human writing patterns, the statistical signatures that detection relies upon diminish. This convergence makes it increasingly difficult for both human users and automated systems to reliably distinguish between human-generated and AI-generated text.15 Knott et al. (2024); Frank et al. (2024) Sadasivan and colleagues provided a theoretical impossibility result indicating that as language models become more sophisticated and better at emulating human text, the performance of even the best-possible detector decreases. This suggests a fundamental constraint on detection accuracy that cannot be overcome through improved algorithms.

Part II: The Systemic Transformation of Scholarly Production

While institutions are hyper-aware of AI use in undergraduate contexts, they are not noticing the far more significant transformation that has already occurred within professional academic research. The evidence suggests that AI-assisted writing has become normative in scholarly publishing within two years of ChatGPT’s release.16 Matsui (2025); Yip et al. (2025)

The most methodologically rigorous study of this phenomenon, published by Kobak and colleagues in Science Advances in July 2025,17 Kobak et al. (2025) analyzed 15.1 million English-language biomedical abstracts indexed by PubMed from 2010–2024. Using an “excess word” methodology inspired by epidemiological studies of COVID-19 mortality, the researchers identified words appearing more frequently than predicted by pre-ChatGPT trends. Their findings indicate that at least 13.5% of 2024 biomedical abstracts—approximately one in seven—were processed with large language models. This lower bound varied significantly by subcategory, reaching 40% for certain disciplines, countries, and journals.

The researchers identified 379 “style words” revealing AI usage, including terms like “delve,” “crucial,” “meticulous,” and “intricate” that showed dramatic frequency increases post-November 2022. The study concluded that “LLMs have had an unprecedented impact on scientific writing in biomedical research, surpassing the effect of major world events such as the COVID pandemic.”18 Kobak et al. (2025)

The peer review process has been similarly affected. Liang and colleagues’ analysis presented at the International Conference on Machine Learning in 202419 Liang et al. (2024) found that between 6.5% and 16.9% of peer reviews at major AI conferences were substantially modified by LLMs—defined as modifications beyond mere spell-checking. The study examined reviews from ICLR 2024, NeurIPS 2023, CoRL 2023, and EMNLP 2023, finding that word frequencies for terms like “commendable” increased 9.8-fold and “meticulous” increased 34.7-fold in ICLR 2024 reviews compared to pre-ChatGPT baselines. Behavioral patterns associated with higher LLM usage included lower reported confidence, submissions closer to deadlines, and decreased likelihood of responding to author rebuttals.

Institutional responses are beginning to acknowledge this reality. In October 2025, arXiv reportedly indicated changes to its computer science category regarding review articles or position papers.20 Sun et al. (2025); Jianghao et al. (2025) This decision was reportedly driven by a significant increase in LLMassisted survey papers, which were often described as little more than annotated bibliographies.21 Liang et al. (2025b); Wang et al. (2024) The platform receives hundreds of such articles monthly and faces challenges in distinguishing quality submissions.

These data points describe a systemic transformation of scholarly production, not an isolated student cheating problem. Policies focused on undergraduate detection while professional practice evolves unchecked represent a category error in institutional response.

13.5%
Of 2024 biomedical abstracts show LLM use
40%
In some fields, countries, and journals
16.9%
Of peer reviews at major AI conferences

Part III: Detection Bias as an Equity Crisis

Beyond unreliability, AI detection tools exhibit systematic bias that transforms their deployment into a civil rights concern.22 Gorichanaz (2023); Giray (2024) This bias is most striking for non-native English speakers, but also extends to neurodivergent students and other populations whose writing patterns diverge from algorithmic expectations of “humanness.”23 Gegg-Harrison and Quarterman (2024); Chambers and Kelley (2025)

The landmark study on this issue, published by Liang and colleagues in Patterns, evaluated seven widely-used GPT detectors against two corpora: 91 TOEFL essays written by non-native English speakers and 88 eighth-grade essays written by native English speakers.24 Liang et al. (2023) While detectors achieved near-perfect accuracy for native speaker essays, they misclassified over half of the non-native English writing samples as AI-generated. Specifically, GPT-based detection tools misclassified 61.3% of non-native English speaker samples.25 Perkins et al. (2024c) The bias was pervasive: 97.80% of TOEFL essays were flagged as AI-generated by at least one detector, and 19.78% were unanimously misclassified by all seven.26 Arango et al. (2024)

The mechanism underlying this bias is algorithmic rather than coincidental.27 Liang et al. (2023) Detection systems operate by measuring “perplexity”— how predictable word choices are.28 Park (2024) Non-native English speakers naturally produce text with less linguistic diversity and more predictable word choices due to limited lexical richness, reduced syntactic diversity, and lower grammatical complexity. The TOEFL essays that were unanimously misclassified showed significantly lower perplexity. The system thus penalizes exactly the linguistic patterns that characterize competent second-language writing.

The civil rights implications are substantial. English Language Learner students—comprising over 10% of the U.S. student population29 Alejandro (2024)—are protected under Title VI of the Civil Rights Act on the basis of race and national origin. False accusations based on AI detector bias could constitute actionable discrimination,30 Jacob et al. (2024); Ovalle (2023); Jiang et al. (2024) a concern highlighted in educational contexts. For international students, such accusations can extend beyond academic penalty, potentially impacting their immigration status and causing significant harm to their academic careers.31 Bannister et al. (2024); Giray (2024) Students often find themselves in a position where they must defend their work against algorithmic accusations.32 Gorichanaz (2023)

Neurodivergent students face parallel vulnerabilities. Research demonstrates that neurodivergent writers, similar to second-language writers, are more likely to receive false positive results from AI text-detection tools.33 Gegg-Harrison and Quarterman (2024); Chambers and Kelley (2025) The mechanism mirrors the non-native speaker problem: autistic students often exhibit writing patterns characterized by rigid style, limited vocabulary, and difficulties expressing thoughts, which may inadvertently resemble algorithmic expectations of AI output.34 França et al. (2024); Chambers and Kelley (2025); Carik et al. (2024)

Detection-based enforcement thus creates a two-tier system based on linguistic privilege. Students whose writing naturally matches algorithmic expectations of “humanness”—typically native English speakers with neurotypical communication patterns—face minimal risk. Those whose legitimate writing diverges from these expectations bear a disproportionate risk of accusation.35 Giray (2024)

Part IV: Transparency Architecture as an Alternative Paradigm

If detection fails both technically and ethically, what alternative exists? The answer lies in building transparency into AI tools themselves—designing systems that make AI contributions visible and bounded rather than attempting to identify them retrospectively through surveillance.

This approach draws on the Transparency by Design framework articulated by Felzmann and colleagues in Science and Engineering Ethics, which adapts principles from GDPR’s Privacy by Design to AI contexts.36 Felzmann et al. (2020) The framework proposes integrating transparency “in a step-by-step manner and as an ex-ante value, not as an afterthought.” For academic AI tools, this means designing systems where AI contributions are inherently traceable rather than requiring post-hoc forensic detection.

Jenni AI was developed with this transparency-first philosophy as a foundational design principle. Rather than positioning AI as a black box that generates content for subsequent human review, the platform’s architecture ensures that human judgment remains central to the writing process while AI assistance remains visible and bounded. Several specific design choices distinguish this approach from bulk generation systems.

Sentence-level autocomplete rather than bulk generation maintains human engagement throughout the writing process. Jenni AI’s autocomplete function produces one sentence at a time, requiring users to review and actively accept each suggestion before proceeding. This contrasts with systems that generate entire paragraphs or documents, which create conditions for wholesale substitution of AI for human thought. The engagement pattern matters: sentence-by-sentence interaction keeps the human writer continuously making decisions about content, structure, and argumentation rather than passively reviewing AI output.

Integrated citation verification through Jenni’s Research Library addresses a core vulnerability of AI-assisted academic text: fabricated or misattributed sources. Jenni AI enables users to import existing research materials via BibTeX files, DOI/PMID/arXiv identifiers, or direct PDF upload. AI suggestions are then grounded in these verified sources, with citations linked to specific documents in the user’s library.37 Walters and Wilder (2023); Spennemann (2025); Mustaffa et al. (2025)

Configurable contribution limits enable institutions and users to enforce appropriate boundaries on AI assistance. Research indicates that AI tools are frequently used and viewed as acceptable for supporting routine tasks such as improving readability, grammar checking, literature reviews, information seeking, editing, and structuring academic papers.38 Shimray and Subaveerapandiyan (2025); Xu (2025); Liao et al. (2024); Arroyo-Machado et al. (2025) This aligns with a broader understanding that while AI can assist in these areas, core analytical and argumentative work is expected to remain human-authored.39 Kwon (2025); Rughiniș et al. (2025)

Multilingual support addresses the equity concerns documented in Part III. Jenni AI operates across more than thirty languages, enabling researchers to work in their native language while producing English-language output. This design choice acknowledges that the goal of academic writing is clear communication of ideas, not demonstration of native English fluency. Non-native speakers using AI for language assistance are not circumventing academic integrity requirements; they are using tools to communicate ideas they have independently developed.40 Filimonovic and Rutzer (2025); Prakash et al. (2025); Amano et al. (2025); Rueda et al. (2025); Sebastian and Baron (2024); Lepp and Sarin (2024)

Process documentation and export flexibility support institutional disclosure requirements. Jenni AI maintains records of AI-assisted versus human-authored content throughout the writing process, aligning with general recommendations for transparency and documentation in academic integrity.41 MLA-CCCC Task Force (2023)

This transparency-first approach repositions AI writing tools as ethical accelerants rather than integrity threats. The tool itself becomes the disclosure mechanism, making AI contributions visible by default rather than requiring after-the-fact forensic investigation.

Part V: Institutional Implementation Pathways

Transitioning from detection to transparency requires coordinated changes in policy, assessment design, and faculty development. Several leading institutions have begun this shift, providing models for broader adoption.

Moving from detection to disclosure requirements represents the foundational policy change. Cornell University’s guidance explicitly states: “We currently do not recommend using current automatic detection algorithms for academic integrity violations using generative AI, given their unreliability.”42 Cornell University (2025) Similarly, Columbia University acknowledges that “AI detection should be treated as a guideline and not a grading metric,” recognizing “risks of misidentification, which can have consequences in the classroom.”43 Columbia University (2025) The University of Pittsburgh has declined to endorse or support any AI detection tools, citing “substantial risk of false positives.”44 University of Pittsburgh (2025a)

The alternative is clear disclosure requirements with procedural protections. COPE’s position statement establishes the framework: authors using AI must be “transparent in disclosing in the Materials and Methods (or similar section) of the paper how the AI tool was used and which tool was used,” while remaining “fully responsible” for all content.45 da Veiga (2025)

Assessment redesign offers more robust integrity evidence than detection. The MLA-CCCC Joint Task Force recommends requiring students to share and comment on version history for written work, using tools that replay writing history to make visible “when text was copied and pasted from another source.”46 MLA-CCCC Task Force (2023)

Interactive oral examinations, where students articulate reasoning in real time, verify understanding regardless of how written products were produced.47 Theobold (2021); Gaines and Burrows (2024); Smith and Tanner (2010) Portfolio approaches with scaffolded reflections document learning processes across time, making wholesale AI substitution impractical.48 Chan and Colloton (2024); Generative AI in Higher Education (2025b); Kong et al. (2024); Pedersen et al. (2018)

The AI Assessment Scale developed by Perkins, Furze, Roe, and MacVaugh provides a five-level framework for calibrating appropriate AI integration across assignment types.49 Perkins et al. (2024b); Perkins et al. (2024a) Level 1 applies to supervised settings testing knowledge retention. Levels 2 and 3 permit AI for ideation and editing respectively. Level 4 allows AI-assisted writing with clear disclosure. Level 5 enables full AI exploration with focus on human curation and oversight. Pilot implementation at one institution showed 5.9% increase in student attainment and 33.3% increase in module passing rates while reducing academic misconduct cases.50 Furze et al. (2024)

Faculty development is prerequisite to effective implementation. AI literacy frameworks, which often include competency areas such as technical understanding, evaluative skills, practical application, and ethical considerations, guide this development.51 UNESCO (2024); Tadimalla and Maher (2024); Chiu et al. (2024) Faculty-specific outcomes include developing AI-informed assignments, creating course-specific policies, and fostering responsible AI usage culture.52 Gupta et al. (2024)

Institutional investment in approved AI tools with transparency features enables consistent implementation, with some universities providing university-managed access and privacy protections for AI tools.53 Ng et al. (2025); Wu et al. (2024) Many universities, including Carnegie Mellon, are engaged in discussions and policy development to manage AI tools, focusing on institutional authentication and data protection.54 Pastva et al. (2024)

The shift from detection to transparency requires cultural change alongside policy revision.55 Zhai et al. (2024); Miah (2024); Laflamme and Bruneault (2025); Lund et al. (2025) The broader academic discourse emphasizes that AI transparency is crucial for intellectual honesty in academic settings.56 Zhang et al. (2025) The goal is neither prohibition nor permissiveness, but the development of reflective practice regarding when, how, and why AI assistance serves legitimate purposes.57 Hysaj et al. (2025); Figueroa et al. (2024)

Conclusion

The evidence presented in this whitepaper leads to an uncomfortable but unavoidable conclusion: institutions cannot detect their way to academic integrity in the age of generative AI.

Detection fails on its own terms. Independent testing shows accuracy rates below 80%,58 Weber-Wulff et al. (2023) false positive rates as high as 50% for widely-used platforms,59 Walters (2023) and systematic bias that flags over 61% of non-native English writing as AI-generated.60 Liang et al. (2023) These are not bugs awaiting patches. The theoretical impossibility result from Sadasivan and colleagues demonstrates that as language models improve, even optimal detectors converge toward random classification.61 Sadasivan et al. (2023) Detection is not temporarily inadequate; it is fundamentally constrained.

Meanwhile, the transformation institutions hoped to prevent has already occurred. At least 13.5% of biomedical abstracts show evidence of AI processing,62 Kobak et al. (2025) 22.5% of computer science papers evidence AI modification,63 Liang et al. (2025a) and up to 16.9% of peer reviews at major conferences are substantially AI-assisted.64 Liang et al. (2024) Policies that focus on undergraduate essays while professional scholarship evolves unchecked address a symptom while ignoring the condition.

The alternative is not permissiveness but transparency. Tools designed with sentence-level engagement, integrated citation verification, and contribution tracking make AI use visible by default rather than detectable after the fact. Combined with disclosure requirements aligned with COPE guidelines65 da Veiga (2025) and process documentation as recommended by the MLA-CCCC Joint Task Force,66 MLA-CCCC Task Force (2023) transparency-first approaches maintain academic integrity without the errors, inequities, and adversarial dynamics of detection-based enforcement.

The question facing institutions is no longer whether to permit AI use. That question has been answered by faculty, researchers, and students who have already integrated these tools into their work. The question now is whether AI use will be pedagogically integrated or covertly substituted.

Institutions that continue investing in detection alone will find themselves administering systems that harm the students they are meant to serve—disproportionately flagging non-native speakers and neurodivergent writers while sophisticated users evade detection entirely. Those that shift toward transparency can shape emerging norms rather than react to them, building frameworks that distinguish legitimate assistance from academic fraud through visibility rather than surveillance.

The path forward requires concrete action. Institutions should audit their current detection tool contracts against the accuracy and bias data presented here. They should develop disclosure policies that treat AI assistance as declarable rather than prohibited. They should invest in faculty development that builds AI literacy across functional, ethical, and rhetorical dimensions. And they should evaluate transparency-oriented tools—platforms designed to make AI contributions traceable—as infrastructure for the new academic environment rather than threats to the old one.

This is not capitulation to technological inevitability. It is recognition that academic integrity has always depended on norms, expectations, and honest disclosure rather than perfect surveillance. The goal remains unchanged: ensuring that credentials reflect genuine learning and that scholarly work represents authentic intellectual contribution. The method must change because the environment has changed. Detection answered a question that no longer admits of reliable answers. Transparency addresses the reality institutions now face.

Declaration of AI Usage

During the preparation of this whitepaper the authors used Jenni AI in order to refine prose and verify citation formatting, as well as improve the clarity and quality of written communication. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

What institutions can do next

  1. 1Audit current detection tool contracts against the accuracy and bias data presented here.
  2. 2Develop disclosure policies that treat AI assistance as declarable rather than prohibited.
  3. 3Invest in faculty development that builds AI literacy.
  4. 4Evaluate transparency-oriented tools as infrastructure, not as threats.

For institutions

Bring a transparency-first writing tool to your department or library.

See Jenni for Teams

For researchers

Join thousands of researchers who trust Jenni for transparent, verifiable work.

Start writing

References

Alejandro, S. (2024). The education balancing act: AI progression, fairness and biases. Policy Perspectives, 31. https://doi.org/10.4079/pp.v31i0.15
Amano, T., Bowker, L., & Burton-Jones, A. (2025). AI-mediated translation presents two possible futures for academic publishing in a multilingual world. PLoS Biology, 23(6). https://doi.org/10.1371/journal.pbio.3003215
Arango, M. C., Hincapié-Otero, M. M., Hardeman, K. N., Shao, B., Starbird, L., & Starbird, C. A. (2024). Special considerations for the use of AI tools by PEERs as a learning and communication aid. Carolina Digital Repository (University of North Carolina at Chapel Hill). https://doi.org/10.17615/m0hg-9687
Arroyo‐Machado, W., Ma, J., Chen, T., Johnson, T. P., Islam, S., Michalegko, L., & Welch, E. W. (2025). Generative AI and academic scientists in US universities: Perception, experience, and adoption intentions. PLoS ONE, 20(8). https://doi.org/10.1371/journal.pone.0330416
Bannister, P., Peñalver, E. A., & Urbieta, A. S. (2024). International students and generative artificial intelligence. Journal of International Students, 14(3), 149–170. https://doi.org/10.32674/jis.v14i3.6277
Bordalejo, B., Pafumi, D., Onuh, F., Khalid, A. K. M. I., Pearce, M. S., & O’Donnell, D. P. (2025). “Scarlet Cloak and the Forest Adventure”: A preliminary study of the impact of AI on commonly used writing tools. International Journal of Educational Technology in Higher Education, 22(1). https://doi.org/10.1186/s41239-025-00505-5
Carik, B., Ping, K., Ding, X., & Rho, E. H. (2024). Exploring large language models through a neurodivergent lens: Use, challenges, community-driven workarounds, and concerns. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2410.06336
Chambers, S., & Kelley, M. C. (2025). The misclassification of autistic writing as AI-generated. In Lecture notes in computer science (pp. 89–103). Springer Science+Business Media. https://doi.org/10.1007/978-3-031-98420-4_7
Chan, C. K. Y., & Colloton, T. (2024). Generative AI in higher education. https://doi.org/10.4324/9781003459026
Cheng, Y., Sadasivan, V. S., Saberi, M., Saha, S., & Feizi, S. (2025). Adversarial paraphrasing: A universal attack for humanizing AI-generated text. https://doi.org/10.48550/ARXIV.2506.07001
Chiu, T. K. F., Ahmad, Z., Ismailov, M., & Sanusi, I. T. (2024). What are artificial intelligence literacy and competency? A comprehensive framework to support them. Computers and Education Open, 6, 100171. https://doi.org/10.1016/j.caeo.2024.100171
Columbia University. (2025). Office of the Provost — Generative AI. https://provost.columbia.edu/content/generative-ai
da Veiga, A. (2025). Ethical guidelines for the use of generative artificial intelligence and artificial intelligence-assisted tools in scholarly publishing: A thematic analysis. Science Editing, 12(1), 28–34. https://doi.org/10.6087/kcse.352
Felzmann, H., Fosch-Villaronga, E., Lutz, C., & Tamò-Larrieux, A. (2020). Towards transparency by design for artificial intelligence. https://doi.org/10.1007/s11948-020-00276-4
Figueroa, B. M., Eaton, S. E., Pethrick, H., Hayden, A., Brennan, R. W., Wiens, J., & McDermott, B. (2024). Academic integrity and artificial intelligence in higher education (HE) contexts: A rapid scoping review. Canadian Perspectives on Academic Integrity, 7(3). https://doi.org/10.55016/ojs/cpai.v7i3.78123
Filimonovic, D., & Rutzer, C. (2025). Generative AI as a linguistic equalizer in global science. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2511.11687
Frank, J. H., Herbert, F., Ricker, J., Schönherr, L., Eisenhofer, T., Fischer, A., Dürmuth, M., & Holz, T. (2024). A representative study on human detection of artificially generated media across countries. https://doi.org/10.1109/sp54263.2024.00159
França, A. B. C., Reategui, E. B., Mintz, J., Meira, R. R., & Motz, R. (2024). Writing analytics and AI for special education: Preliminary results on students with autism spectrum disorder. In Communications in computer and information science (pp. 192–199). Springer Science+Business Media. https://doi.org/10.1007/978-3-031-64312-5_23
Furze, L., Perkins, M., Roe, J., & MacVaugh, J. (2024). The AI assessment scale (AIAS) in action: A pilot implementation of GenAI-supported assessment. Australasian Journal of Educational Technology. https://doi.org/10.14742/ajet.9434
Gaines, T., & Burrows, N. L. (2024). Exploring alternative assessments during COVID: Instructor experiences using oral exams. Frontiers in Education, 9. https://doi.org/10.3389/feduc.2024.1379886
Gegg-Harrison, W., & Quarterman, C. (2024). AI detection's high false positive rates and the psychological and material impacts on students. In Advances in educational marketing, administration, and leadership book series (pp. 199–219). IGI Global. https://doi.org/10.4018/979-8-3693-0240-8.ch011
Generative AI in higher education. (2025). Edward Elgar Publishing. https://doi.org/10.4337/9781035326020
Giray, L. (2024). The problem with false positives: AI detection unfairly accuses scholars of AI plagiarism. The Serials Librarian, 1–9. https://doi.org/10.1080/0361526x.2024.2433256
Gorichanaz, T. (2023). Accused: How students respond to allegations of using ChatGPT on assessments. Learning Research and Practice, 9(2), 183–196. https://doi.org/10.1080/23735082.2023.2254787
Gupta, A., Atef, Y., Mills, A., & Bali, M. (2024). Assistant, parrot, or colonizing loudspeaker? ChatGPT metaphors for developing critical AI literacies. Open Praxis, 16(1), 37–53. https://doi.org/10.55982/openpraxis.16.1.631
Hysaj, A., Dean, B. A., & Freeman, M. (2025). Exploring the purposes and uses of generative artificial intelligence tools in academic writing for multicultural students. Higher Education Research & Development, 1–15. https://doi.org/10.1080/07294360.2025.2488862
Is AI-generated content actually detectable? (2025). University of Maryland. https://cmns.umd.edu/news-events/news/ai-generated-content-actually-detectable
Jabarian, B., & Imas, A. (2025). Artificial writing and automated detection. https://doi.org/10.2139/ssrn.5407424
Jacob, S., Tate, T., & Warschauer, M. (2024). Emergent AI-assisted discourse: A case study of a second language writer authoring with ChatGPT. Journal of China Computer-Assisted Language Learning. https://doi.org/10.1515/jccall-2024-0011
Jiang, Y., Hao, J., Fauss, M., & Li, C. (2024). Detecting ChatGPT-generated essays in a large-scale writing assessment: Is there a bias against non-native English speakers? Computers & Education, 217, 105070. https://doi.org/10.1016/j.compedu.2024.105070
Jianghao, L., Rong, S., Jiachen, Z., Xi, Y., Yong, Y., & Zhang, W. (2025). Stop DDoS attacking the research community with AI-generated survey papers. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2510.09686
Knott, A., Pedreschi, D., Jitsuzumi, T., Leavy, S., Eyers, D., Chakraborti, T., Trotman, A., Sundareswaran, S., Baeza‐Yates, R., Biecek, P., Weller, A., Teal, P. D., Basu, S., Haklıdır, M., Morini, V., Russell, S., & Bengio, Y. (2024). AI content detection in the emerging information ecosystem: New obligations for media and tech companies. Ethics and Information Technology, 26(4). https://doi.org/10.1007/s10676-024-09795-1
Kobak, D., González-Márquez, R., Horvát, E. Á., & Lause, J. (2025). Delving into LLM-assisted writing in biomedical publications through excess vocabulary. Science Advances, 11(27). https://doi.org/10.1126/sciadv.adt3813
Kong, S. C., Lee, J. C., & Tsang, O. (2024). A pedagogical design for self-regulated learning in academic writing using text-based generative artificial intelligence tools: 6-P pedagogy of plan, prompt, preview, produce, peer-review, portfolio-tracking. Research and Practice in Technology Enhanced Learning, 19, 30. https://doi.org/10.58459/rptel.2024.19030
Kwon, D. (2025). Is it OK for AI to write science papers? Nature survey shows researchers are split. Nature, 641(8063), 574–578. https://doi.org/10.1038/d41586-025-01463-8
Laflamme, A. S., & Bruneault, F. (2025). Redefining academic integrity in the age of generative artificial intelligence: The essential contribution of artificial intelligence ethics. Journal of Scholarly Publishing, 56(2), 481–509. https://doi.org/10.3138/jsp-2024-1125
Latona, G. R., Ribeiro, M. H., Davidson, T. R., Veselovsky, V., & West, R. (2024). The AI review lottery: Widespread AI-assisted peer reviews boost paper scores and acceptance rates. https://doi.org/10.48550/ARXIV.2405.02150
Lepp, H., & Sarin, P. (2024). A global AI community requires language-diverse publishing. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2408.14772
Liang, W., Izzo, Z., Zhang, Y., Lepp, H., Cao, H., Zhao, X., Chen, L., Ye, H., Liu, S., Huang, Z., McFarland, D. A., & Zou, J. (2024). Monitoring AI-modified content at scale: A case study on the impact of ChatGPT on AI conference peer reviews. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2403.07183
Liang, W., Yüksekgönül, M., Mao, Y., Wu, E. Q., & Zou, J. (2023). GPT detectors are biased against non-native English writers. Patterns, 4(7), 100779. https://doi.org/10.1016/j.patter.2023.100779
Liang, W., Zhang, Y., Wu, Z., Lepp, H., Ji, W., Zhao, X., Cao, H., Liu, S., He, S., Huang, Z., Yang, D., Potts, C., Manning, C. D., & Zou, J. (2025a). Quantifying large language model usage in scientific papers. Nature Human Behaviour. https://doi.org/10.1038/s41562-025-02273-8
Liang, X., Yang, J., Wang, Y., Tang, C., Zheng, Z., Niu, S., Song, S., Wang, H., Tang, B., Xiong, F., Mao, K., & Li, Z. (2025). SurveyX: Academic survey automation via large language models. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2502.14776
Liao, Z., Antoniak, M., Cheong, I., Cheng, E. Y. Y., Lee, A. H., Lo, K., Chang, J. C., & Zhang, A. X. (2024). LLMs as research tools: A large scale survey of researchers' usage and perceptions. https://doi.org/10.48550/ARXIV.2411.05025
Liu, J. Q. J., Hui, K. T., Zoubi, F. A., Zhou, Z. Z. X., Samartzis, D., Yu, C. C. H., Chang, J. R., & Wong, A. Y. L. (2024). The great detectives: Humans versus AI detectors in catching large language model-generated medical writing. International Journal for Educational Integrity, 20(1). https://doi.org/10.1007/s40979-024-00155-6
Lund, B., Orhan, Z., Mannuru, N. R., Bevara, R. V. K., Porter, B., Vinaih, M. K., & Bhaskara, P. (2025). Standards, frameworks, and legislation for artificial intelligence (AI) transparency. AI and Ethics, 5(4), 3639–3655. https://doi.org/10.1007/s43681-025-00661-4
Matsui, K. (2025). Delving into PubMed records: How AI-influenced vocabulary has transformed medical writing since ChatGPT. Perspectives on Medical Education, 14(1), 882–890. https://doi.org/10.5334/pme.1929
Miah, M. (2024). Unveiling the evolutionary impact of artificial intelligence on the workforce. Informatica Economica, 28, 39–58. https://doi.org/10.24818/issn14531305/28.1.2024.04
MLA-CCCC Joint Task Force on Writing and AI. (2023). Academic integrity and assignment design. https://aiandwriting.hcommons.org/2023/09/22/academic-integrity-and-assignment-design/
Mustaffa, N. E., Lai, K. H., Preece, C., & Wong, F. Y. (2025). A bibliometric review of large language model hallucination. International Journal of Research and Innovation in Social Science, 5025–5037. https://doi.org/10.47772/ijriss.2025.909000409
Ng, B., Liu, J., Tong, X., Ye, K., Yenne, G., Chandrasekaran, V., & Li, J. (2025). Analyzing security and privacy challenges in generative AI usage guidelines for higher education. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2506.20463
Ovalle, A. (2023). Decoding the digital fuku: Deciphering colonial legacies to critically assess ChatGPT in Dominican education. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2310.17533
Park, S. (2024). AI chatbots and linguistic injustice. Journal of Universal Language, 25(1), 99–119. https://doi.org/10.22425/jul.2024.25.1.99
Pastva, J., Jebbia, D., Reilly, M., & Werlinich, A. (2024). The implementation of Keenious at Carnegie Mellon University. Journal of eScience Librarianship, 13(1). https://doi.org/10.7191/jeslib.800
Pedersen, A. Y., Caviglia, F., Kay, J., & Luckin, R. (2018). Researcher or fellow citizen?: Looking for a role model in the humanities. Research Portal Denmark, 2, 945–948.
Perkins, M., Furze, L., Roe, J., & MacVaugh, J. (2024). The artificial intelligence assessment scale (AIAS): A framework for ethical integration of generative AI in educational assessment. Journal of University Teaching and Learning Practice, 21(6). https://doi.org/10.53761/q3azde36
Perkins, M., Roe, J., & Furze, L. (2024). The AI assessment scale revisited: A framework for educational assessment. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2412.09029
Perkins, M., Roe, J., Vu, B., Postma, D., Hickerson, D., McGaughran, J., & Khuat, H. Q. (2024). GenAI detection tools, adversarial techniques and implications for inclusivity in higher education. Cornell University. https://doi.org/10.48550/arxiv.2403.19148
Prakash, A., Aggarwal, S., Varghese, J. J., & Varghese, J. (2025). Writing without borders: AI and cross-cultural convergence in academic writing quality. Humanities and Social Sciences Communications, 12(1). https://doi.org/10.1057/s41599-025-05484-6
Rueda, J., Ivanković, V., & Blunden, C. (2025). A cautiously optimistic view on large language model use against linguistic injustice in academia. Nature Machine Intelligence, 7(11), 1773–1774. https://doi.org/10.1038/s42256-025-01138-z
Rughiniș, C., Vulpe, S., Țurcanu, D., & Rughiniș, R. (2025). AI at the knowledge gates: Institutional policies and hybrid configurations in universities and publishers. Frontiers in Computer Science, 7. https://doi.org/10.3389/fcomp.2025.1608276
Sadasivan, V. S., Kumar, A., Balasubramanian, S., Wang, W., & Feizi, S. (2023). Can AI-generated text be reliably detected? arXiv (Cornell University). https://doi.org/10.48550/arxiv.2303.11156
Sebastian, F., & Baron, R. (2024). Artificial intelligence: What the future holds for multilingual authors and editing professionals. Science Editor. https://doi.org/10.36591/se-d-4702-01
Shimray, S. R., & Subaveerapandiyan, A. (2025). Artificial intelligence in academic writing and research: Adoption and effectiveness. Open Information Science, 9(1). https://doi.org/10.1515/opis-2025-0026
Smith, J. I., & Tanner, K. D. (2010). The problem of revealing how students think: Concept inventories and beyond. CBE—Life Sciences Education, 9(1), 1–5. https://doi.org/10.1187/cbe.09-12-0094
Spennemann, D. (2025). The origins and veracity of references ‘cited’ by generative artificial intelligence applications: Implications for the quality of responses. Publications, 13, 12. https://doi.org/10.3390/publications13010012
Sun, Z., Zhu, X., Zhou, X., Tong, X., Wang, S., Fu, J., Li, G., Liu, Z., & Wu, F. (2025). SurveyBench: Can LLM(-agents) write academic surveys that align with reader needs? arXiv (Cornell University). https://doi.org/10.48550/arxiv.2510.03120
Tadimalla, S. Y., & Maher, M. L. (2024). AI literacy for all: Adjustable interdisciplinary socio-technical curriculum. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2409.10552
Theobold, A. S. (2021). Oral exams: A more meaningful assessment of students’ understanding. Journal of Statistics and Data Science Education, 29(2), 156–159. https://doi.org/10.1080/26939169.2021.1914527
UNESCO. (2024). AI competency framework for teachers. https://doi.org/10.54675/zjte2084
University of Pittsburgh. (2025). Generative AI: Encouraging academic integrity. https://teaching.pitt.edu/resources/encouraging-academic-integrity/
Wakjira, T. G., Tijani, I. A., Alam, M. S., Mashal, M., & Hasan, K. (2025). Can we trust AI content detection tools for critical decision-making? Information, 16(10), 904. https://doi.org/10.3390/info16100904
Walters, W. H. (2023). The effectiveness of software designed to detect AI-generated writing: A comparison of 16 AI text detectors. Open Information Science, 7(1). https://doi.org/10.1515/opis-2022-0158
Walters, W. H., & Wilder, E. I. (2023). Fabrication and errors in the bibliographic citations generated by ChatGPT. Scientific Reports, 13(1). https://doi.org/10.1038/s41598-023-41032-5
Wang, Y., Guo, Q., Yao, W., Zhang, H., Zhang, X., Wu, Z., Zhang, M., Dai, X., Zhang, M., Wen, Q., Ye, W., Zhang, S., & Zhang, Y. (2024). Autosurvey: Large language models can automatically write surveys. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2406.10252
Weber-Wulff, D., Anohina-Naumeca, A., Bjelobaba, S., Foltýnek, T., Guerrero-Dib, J., Popoola, O., Šigut, P., & Waddington, L. (2023). Testing of detection tools for AI-generated text. International Journal for Educational Integrity, 19(1). https://doi.org/10.1007/s40979-023-00146-z
Wu, C., Zhang, H., & Carroll, J. M. (2024). AI governance in higher education: Case studies of guidance at Big Ten universities. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2409.02017
Xu, Z. (2025). Patterns and purposes: A cross-journal analysis of AI tool usage in academic writing. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2502.00632
Yip, R., Sun, Y. J., Bassuk, A. G., & Mahajan, V. B. (2025). Artificial intelligence’s contribution to biomedical literature search: Revolutionizing or complicating? PLOS Digital Health, 4. https://doi.org/10.1371/journal.pdig.0000849
Zhai, C., Wibowo, S., & Li, L. D. (2024). The effects of over-reliance on AI dialogue systems on students’ cognitive abilities: A systematic review. https://doi.org/10.1186/s40561-024-00316-7
Zhang, X., Zhang, J., & Oubibi, M. (2025). Effects of Chinese college students’ academic integrity perceptions on the intention to disclose AI-generated outputs: A moderated mediation model. Journal of Academic Ethics, 23(4), 1709–1728. https://doi.org/10.1007/s10805-025-09622-2

Cite this paper

Spezowka, Alex. Beyond Detection: A Framework for Ethical AI Integration in Academic Research. Jenni Research, 2026. jenni.ai/whitepaper.

Trusted by Universities and businesses across the world

google logo
penn logo
cambridge university
aston university logo

Trusted by Universities and businesses across the world

google logo
penn logo
cambridge university
aston university logo

Trusted by Universities and businesses across the world

google logo
penn logo
cambridge university
aston university logo
INTEGRITY BY DESIGN

Built for trust, not just speed

Every piece of work created with Jenni carries built-in transparency and verifiability.

Write with confidence

Verify text directly against cited sources to ensure your claims accurately represent the literature.

Claim confidence

The claim confidence analysis addressed issues of redundant, weak, or missing citations, alongside instances of contradiction in citation arguments.

Misrepresented

Contradicted

3

Unsupported

4

Weakly Supported

2

Overstated

Unverifiable

Catch issues early and prevent rounds of feedback

Check your claims against existing research, ensuring your arguments are defensible.

The home for your research writing

No more wrestling with app integrations, write your next paper in one place.

Data Privacy

Your drafts stay private. Jenni does not use your writing to train models, and our upstream AI providers are contractually prohibited from training on Jenni user data.

AI Declaration

Insert a compliance statement that aligns with disclosure standards from frameworks like the OECD AI Principles and UNESCO's Ethics of AI.

Built for Collaboration

Write, edit, and cite with your co-authors in one, easy to use document editor. Avoid the pain of versioning with seamless collaboration.

I used Jenni to craft synopses for my PhD applications, focusing on digitalization, supply chain optimization, and resilience. The tool helped me change my research ideas into concise, compelling proposals that caught the attention of many readers. 

Rebecca Nakalanda | Lecturer, Makerere University Business School

Ready to write with confidence?

Join thousands of researchers who trust Jenni for transparent, verifiable work.

Start for free

No credit card required

Cancel anytime

Over 8m

Academics worldwide

5.2 hours saved

On average per paper

Over 15m

Papers written on Jenni

Ready to write with confidence?

Join thousands of researchers who trust Jenni for transparent, verifiable work.

Start for free

No credit card required

Cancel anytime

Over 8m

Academics worldwide

5.2 hours saved

On average per paper

Over 15m

Papers written on Jenni

Ready to write with confidence?

Join thousands of researchers who trust Jenni for transparent, verifiable work.

Start for free

No credit card required

Cancel anytime

Over 8m

Academics worldwide

5.2 hours saved

On average per paper

Over 15m

Papers written on Jenni