The Creation of Bad Students: AI Detection for Non-Native English Speakers

January 21, 2025

The Creation of Bad Students: AI Detection for Non-Native English Speakers

Creating the “Bad” Student

As a Graduate Student Instructor, I hate proctoring midterms. In a big room, everyone was seated nervously staring at their exams. I was instructed to watch the students: vigilant, my job was to police them. Someone raised their hand and said, “Excuse me, can I pick up my pencil? It fell.” I was confused by the question but realized they needed my approval. They carefully picked up the pencil, trying not to appear as if they were taking advantage of the situation to look around. It was a performance of self-control. Their discipline was complete, and a “good” student was born. I had accomplished my task of disciplining; I was the big eye.

Foucault, in his book Discipline and Punish, said, “Discipline is a political anatomy of detail...To establish the useful communications, to interrupt those which are not, to be able at every moment to watch the conduct of each one, to appreciate it, to sanction it, to measure the qualities and the merits. A protocol to know, to dominate, and to use” (Foucault, 1977). There is a list of details in educational institutions subject to inspection to fulfill the morality of a good student. In these details lies what is allowed for a “normal” student, but at the same time, the description of “abnormal.” Everyone is classified within the hierarchy of the educational institution. Education is not being developed to accommodate diverse students; instead, diverse students are homogenized to fit the institution. As expected, the closer their life experiences and opportunities are to this list of details, the easier it is to conform and navigate the educational institution.

A Bad Student Uses AIs

Of course, vigilance evolves with time and technology. For example, computational models that highlight plagiarism are used at UC Berkeley every semester. Plagiarism is defined as “the use of intellectual material produced by another person without acknowledging its source” (Academic Honesty | College Writing Programs, s. f.). An exemplary punishment is expected if accused of plagiarism at UC Berkeley, as mentioned on Quora by a UC Berkeley student seeking advice after confessing plagiarism: “If you did plagiarize, you violated what is, for many academics, THE most important code of conduct…You shouldn’t expect anything less than what is listed as the first offense punishment for plagiarizing” (Quora, s. f.). This is an example of how academia disciplines and creates the “bad” students.

Now, the debate about what constitutes plagiarism and originality has been reignited with the emergence of artificial intelligence (AI) based on large language models such as ChatGPT (OpenAI, 2023). Plagiarism at UC Berkeley includes “The use of generative AI, such as ChatGPT or Grammarly, to write parts of an assignment without disclosing it, and the use of auto-translation, such as Google Translate, to translate passages from another language into English without disclosing it” (Academic Honesty | College Writing Programs, s. f.). Thus, the current debate around plagiarism focuses on how to detect and punish the use of AIs by students.

Although it seems that the academic community has already made its verdict about plagiarism, the definition of originality itself is unclear. “Copyright is a constant social negotiation, forged tenaciously, revised countless times, and imperfect in each of its incarnations,” said Jonathan Lethem in his essay The Ecstasy of Influence (Lethem, 2007). As social beings, no one learns anything from scratch, and it is impossible to trace a single, unique origin of ideas. Plagiarism, then, is just a moral construction. Moreover, it has been shown that when we assign ideas to specific authors, we tend to do so following our own cultural biases. For example, it often seems that all we know about the organisms of the Andes Cordillera in South America comes from a single European male who visited those places on a colonial expedition (Asymmetric Ecologies in Europe and South America around 1800, 2022). Currently, these debates about the complexity of defining originality include AIs. Is ChatGPT an automated version of multiple plagiarisms combined, or is appropriate/holistic citation simply an impossible task to accomplish? (see ChatGPT’s answer below).

As a biologist, I understand and agree that directing people to important sources of information is crucial to tracing a path of thoughts and evidence. However, there is a landscape of pedagogical strategies outside of punishment, so why do we insist on a paradigm of education by fear? As Foucault describes, discipline helps the advance of capitalism because “it is a matter of concentrating the forces of production, of obtaining from them the maximum of advantages and neutralizing their disadvantages.” Thus, student discipline increases efficiency and defines the accepted student and future worker, which fits perfectly with the maintenance and advancement of a neoliberal education (Mintz, 2021). The call for social justice in education requires a novel model for education, different from the business perspective of neoliberalism, and thus the implementation of novel pedagogical strategies and the development of creative uses of AIs in education.

Are Non-Native English Speakers Bad Students?

After centuries of British colonial activities and U.S. imperialism, English became the dominant language of academia (Gordin, 2015). This has caused non-native English speakers (NNES) to suffer a series of disadvantages when writing, reading, publishing, and understanding science and other academic work published only in English (Amano et al., 2016; Ramírez-Castañeda, 2020). Furthermore, the mandatory use of English in academia has increased the dependency of NNES researchers on native English speakers, which exacerbates global inequalities (Dei & Kempf, 2006; Mahdjoub et al., 2023). AIs are a crucial tool widely used by NNES to improve grammar, translate, and, in general, to partially overcome the individual language challenges we constantly face in academia (Steigerwald et al., 2022). As a NNES myself, I rely heavily on ChatGPT as a useful tool to review my writing and reduce my dependency on native speakers, including when writing this piece. However, under the standards being proposed, this would be considered punishable.

There has been increased conversation about using AI detectors to detect the use of AIs in academia and education, thus creating a new label of “bad” students in academia. AI detectors are large language models created to highlight “suspicious” characteristics of a particular text. This detection is intended to find a fingerprint of another AI. A recent paper by Liang et al. (2023) has shown that the current versions of AI detectors are biased against texts written by NNES. This paper showed that more than 19% of the text produced by NNES was unanimously misclassified as AI creations by seven different AI detectors because they had a more constrained use of the English language, and 97% of these texts were labeled as AI creations by at least one AI detector (Liang et al., 2023). Although improved versions of AI detectors may arise, I do not agree that starting an arms race between AI creators and AI detectors is what we need to improve education. Instead, I believe the use of AI is an important tool to overcome language barriers and advance multilingual academia where more NNES can fairly and fully participate.

Are non-native English students bad students? We cannot escape from language vigilance. We communicate differently than hegemonic native English students (primarily European-descendence US/UK English), an aspect that is constantly framed as negative in academia. For example, different accents were “corrected” in the mandatory international GSI training that I attended in my first semester at UC Berkeley. On the other hand, when we seek help, for example, by using AI technologies to improve our writing, we are punished or asked to constantly disclose our cultural identity–a fact that shouldn’t be a problem, but it is, in this biased world (Gaddis, 2017; Kozlowski et al., 2022). We are trapped. It seems that in academia, NNES will always be at a disadvantage. But we are not the only ones who lose from this, minds that have not been fully disciplined are the ones that propose and live in alternative worlds, worlds where diversity is a value, and education is an experience of care.

Conclusion for/from a Bad Student

The current educational paradigm perpetuates a cycle of surveillance and punishment, conflating “originality” with a rigid, culturally biased notion of intellectual property. At UC Berkeley, the use of AI in academia has been labeled as dishonest behavior, unfairly penalizing non-native English speakers (NNES). Yet, for NNES, AI tools are essential for overcoming language barriers and reducing the global and social inequities reinforced by the dominance of the English language. Educational institutions must shift from punishing AI use to embracing creative ways to integrate these tools into the learning process. By doing so, we foster an environment that celebrates diverse perspectives and ideas, moving towards an educational model that prioritizes creativity over control.

P.S. I asked ChatGPT: "Is ChatGPT an automated version of multiple plagiarisms combined, or is appropriate/holistic citation simply an impossible task to accomplish?" It responded:

“The question highlights the ethical and technical challenges of AI like ChatGPT in academic contexts. It suggests AI might unintentionally resemble plagiarism by synthesizing content without attribution, but AI doesn't copy directly; it generates new text based on learned patterns. Proper citation by AI is difficult, as models don't track specific sources. This raises concerns about using AI responsibly in academia while maintaining originality and ethical standards.”

Notes on the References

  • Ironically, I am including citations in this blog. This is less about intellectual property and more to encourage further exploration of these topics.

  • I used ChatGPT v3.5 to create the post image, correct the grammar, and translate some quotes. I hope that by acknowledging the use of AI, I can help normalize it as a tool, particularly for non-native English speakers navigating the imposed dominance of English in academia. This statement aims to resist the self-disciplining practices being developed to control and punish AI use in education, rather than contribute to them.

References

  1. Academic Honesty | College Writing Programs. (s. f.). Recuperado 13 de agosto de 2024, de. https://writing.berkeley.edu/academic-honesty
  2. Amano, T., González-Varo, J. P., & Sutherland, W. J. (2016). Languages Are Still a Major Barrier to Global Science. PLoS biology, 14(12), e2000933. https://doi.org/10.1371/journal.pbio.2000933
  3. Asymmetric Ecologies in Europe and South America around 1800. (2022). De Gruyter. https://doi.org/10.1515/9783110733211
  4. Dei, G. J. S., & Kempf, A. (2006). Anti-colonialism and education: The politics of resistance. Sense Publishers. https://books.google.es/books?hl=es&lr=&id=HyupVZD5SzwC&oi=fnd&pg=PA25&dq=neocolonialism+scientific&ots=V4jzVjjU01&sig=qazgYr4oNRMlD7514pyXnuOQN68#v=onepage&q=neocolonialism scientific&f=false
  5. Foucault, M. (1977). Discipline and Punish: The Birth of the Prison. Vintage Books.
  6. Gaddis, S. M. (2017). Racial/Ethnic Perceptions from Hispanic Names: Selecting Names to Test for Discrimination. Socius, 3, 2378023117737193. https://doi.org/10.1177/2378023117737193
  7. Gordin, M. D. (2015). Scientific Babel. En Scientific Babel. https://doi.org/10.7208/chicago/9780226000329.001.0001
  8. I plagiarized unintentionally and was caught. The professor reported me to the university. I’m currently enrolled at UC Berkeley. What ca... (s. f.). Quora. Recuperado 13 de agosto de 2024, de https://www.quora.com/I-plagiarized-unintentionally-and-was-caught-The-professor-reported-me-to-the-university-I-m-currently-enrolled-at-UC-Berkeley-What-can-I-expect-as-consequences/answer/Phill-Alexander-2
  9. Kozlowski, D., Murray, D. S., Bell, A., Hulsey, W., Larivière, V., Monroe-White, T., & Sugimoto, C. R. (2022). Avoiding bias when inferring race using name-based approaches. PLoS ONE, 17(3), e0264270. https://doi.org/10.1371/journal.pone.0264270
  10. Lethem, J. (2007). The Ecstasy of Influence: From Cellophane Bricks: A Life in Visual Culture, which was published last month by ZE Books. The museums of New York in the Seventies and Eighties were as much the…. Harper’s Magazine, February 2007. https://harpers.org/archive/2007/02/the-ecstasy-of-influence/
  11. Liang, W., Yuksekgonul, M., Mao, Y., Wu, E., & Zou, J. (2023). GPT detectors are biased against non-native English writers (arXiv:2304.02819). arXiv. https://doi.org/10.48550/arXiv.2304.02819
  12. Mahdjoub, H., Baaloudj, A., Chaib, S., Ramírez-Castañeda, V., Contreras, L. A. B., González-Tokman, D., Villada-Bedoya, S., Rocha-Ortega, M., Córdoba-Aguilar, A., & Khelifa, R. (2023). Benefits and geography of international collaboration for PhD students in biology from four global south countries. Frontiers in Education, 8. https://www.frontiersin.org/articles/10.3389/feduc.2023.1281219
  13. Mintz, B. (2021). Neoliberalism and the Crisis in Higher Education: The Cost of Ideology. The American Journal of Economics and Sociology, 80(1), 79-112. https://doi.org/10.1111/ajes.12370
  14. OpenAI. (2023). ChatGPT (Versión August 2023) [Software]. https://chat.openai.com/
  15. Ramírez-Castañeda, V. (2020). Disadvantages in preparing and publishing scientific papers caused by the dominance of the English language in science: The case of Colombian researchers in biological sciences. PLOS ONE, 15(9), e0238372. https://doi.org/10.1371/journal.pone.0238372
  16. Steigerwald, E., Ramírez-Castañeda, V., Brandt, D. Y. C., Báldi, A., Shapiro, J. T., Bowker, L., & Tarvin, R. D. (2022). Overcoming Language Barriers in Academia: Machine Translation Tools and a Vision for a Multilingual Future. BioScience, 72(10), 988-998. https://doi.org/10.1093/biosci/biac062