We’re in a writing class working on drafts for the first project. I open up a student’s draft in Google Docs, noticing that this time, I am logged in as an Anonymous Capybara. The student is typing, pauses, and suddenly there is an entire paragraph of pasted text from a source that appears to be unknown. Is this the work of generative artificial intelligence (GenAI)? Is this from another source on the web? Did the student work in a different document and is now pasting their ideas into this one? I navigate to the history of the Doc and review the previous versions of writing. The student finally notices that I am in the Doc, and approaches me at my desk, moved to preemptively justify that the work that they pasted into the Doc was their own writing. They assure me that “everything would be cited” and that they didn’t “use AI.”
This short anecdote, or a version of it, is one that I have heard repeated from numerous writing instructors. It reflects the in-the-loop writing practices that influence how surveillance impacts composing in different spaces (refer to Salvo & Sherrill, 2025; Banville & Sugg, 2021). While this classroom interaction may appear mundane, it reflects a much larger network of monitoring practices embedded within contemporary educational technologies and institutional expectations. Surveillance studies scholars describe these interconnected systems of observation, data collection, and behavioral management as a “surveillance assemblage” (Haggerty & Ericson, 2000). For the purposes of this blog, surveillance may be defined as “the collection of both visible and invisible data/information derived from those being observed, suggesting an application of power over the observed audience, who are often not informed of such collection” (Banville, 2023, p. 32). The uptick of surveillance technologies such as generative artificial intelligence (GenAI) usage within classrooms across the globe has contributed to the institutional and rhetorical conditions that influence trust, credibility, and critical thinking. Students are operating under hyper-surveillance by their instructors, while instructors are feeling pressured to decipher what is or is not student writing. Such pressure to surveil students may come from institutional administrators, program directors, or other leaders within higher education. However, there is also individual instructor pressure; that is, there is a pedagogical expectation that each instructor takes a stance on surveillance technologies such as GenAI, create a policy, and figure out the “best practices.”
This pressure to figure out what to do, how to act, and how to navigate approaching student work and expectations has, in some contexts, seemingly led to composing practices that are filtered, trust that is absent, and students who believe that their every movement is surveilled (and they’re right). Instructors across the globe have struggled with the influx of GenAI usage, especially, within the college classroom space (refer to Neumann & Gerstl-Pepin, 2025). Such “hype” and perhaps oversaturation has led to conversations about misinformation, energy consumption, academic integrity, data privacy, budgetary constraints, and intellectual property rights. Writing instructors in particular have been grappling with the tensions between institutional goals of academic integrity and active participation and “the unintended consequences of fostering performative behaviours and eroding relational trust” (Calderwood, 2025, p. 1). There has been an increase in plagiarism detectors, lock-down browsers, biometric attendance, and increased surveillance through Learning Management Systems (LMS)—all of which contribute to encouraging compliance from students and behavioral control, rather than encouraging authentic engagement. Ultimately, the aforementioned surveillance mechanisms promote compliance over genuine learning, which in turn contributes to the instructor-student dynamic and their composing processes. This warped reality informs not just composing processes, but it is indicative of a surveillance assemblage (refer to Haggerty & Ericson, 2000), of which every party is subjected to surveillance—to varying degrees and for different reasons. Surveillance assemblages operate by “abstracting human bodies from their territorial settings and separating them into a series of discrete flows. These flows are then reassembled into distinct ‘data doubles’ which can be scrutinized and targeted for intervention” (Haggerty & Ericson, 2000, p. 605). The surveillance assemblage in higher education occurs mostly through educational technologies, relying on digital systems such as LMS, proctoring tools, and security cameras to collect and aggregate student data into a continuous, trackable “data double.” Edwards writes how data doubles are “vast archives of personalized information [that]circulate absent the territorialized body, traveling via infrastructures subject to sorting, categorization, and other surveillance techniques. In other words, a multiplicity of circulatory flows is always being produced through everyday acts of composing, flows that can be aggregated, duplicated, sorted, and recirculated” (2021, p. 79). Despite these challenges, some instructors employ strategies of resistance to surveillance technologies and strategies by “prioritising human connection and relational pedagogy to counteract the dehumanising effects of constant monitoring” (Calderwood, 2025, p. 1). We cannot escape the surveillance-in-the-loop processes that involve varying power dynamics and complex entanglements and decision-making; but perhaps we as instructors can try to re-center the human and relational aspects of writing.
Each instructor, or campuses broadly speaking, have had to decide how to approach the conflicting conversations about GenAI usage and the place of such surveillance technologies in classrooms. On a larger scale, universities have decided to incorporate or actively refuse GenAI, effectively making the choice for instructors and taking away any academic freedom (refer to CSU, 2026). Wherever surveillance exists, power dynamics are also involved. Though it is “easy” to focus on the instructor-student surveillance relationship, there is also a larger surveillance assemblage, one where instructors are surveilled by university administrators → who are surveilled by BigTech → who also surveil one another, including students and staff. Though surveillance is colloquially viewed as “top-down” or “panoptic,” it is important to note that such power dynamics occur multi-directionally: for example, laterally by peers, or even surveilling those in positions of power (also known as sousveillance).
There are several implications as to how power and surveillance practices can inform the ways in which students compose. Such surveillance of not just students, but also faculty, staff, and administrators, blends “private-sector data collection with public educational goals, often prioritizing control, efficiency, and safety over privacy, creating a ‘surveillant assemblage’ that monitors students to shape behavior and automate compliance” (Crooks, 2022, n.p.). Most common is the surveillance-in-the-loop writing and composing process, of which student work is monitored to ensure authenticity and integrity, occasionally shaping the way students behave (depending on the level of awareness that they are being surveilled, how, and through what form), as well as effectively widening the gap of trust between students and their instructors. In this surveillance loop, no one is trusted and everyone (and their writing) is under suspicion. The surveillance-in-the-loop writing process is a microcosm of a larger surveillance apparatus, one that we are all participants in. In the United States in particular, we are tracked digitally and physically by governmental and BigTech companies (i.e. Google, Meta, Amazon, and so forth). We are tracked through browser cookies and pixels, IP address monitoring (especially from employers), through social media habits (including likes, shares, seconds and number of times viewed), GPS, cell tower triangulation, Wi-Fi networks, Bluetooth, mobile apps that share location, and much more additional personal data. Without adequate privacy regulations in place, and without placing responsibility for privacy protection on the individual, we are effectively stuck in this surveillance loop.
At the most mundane level, surveillance occurs to navigate everyday life, whether through online browsing, border crossing, providing health information, or going grocery shopping. The privacy problems associated with the surveillance-in-the-loop composing process deserve serious intellectual consideration. As Crooks (2022) notes, “data captured, aggregated, and analyzed by contemporary computational systems…produce new opportunities to sort people, a power that rests in the hands of technologically sophisticated corporate and state actors exclusively” (n.p.). Surveillance technologies harm vulnerable populations, and though leveraged as inevitable and ubiquitous, can be resisted, challenged, and critically questioned. The tools we use in the classroom space do not need to surveil our students or their composing processes. It is time to break the loop.
We’re in a writing class working on drafts for the first project. I open up a student’s draft in Google Docs, noticing that this time, I am logged in as an Anonymous Capybara. The student is typing, pauses, and suddenly there is an entire paragraph of pasted text from a source that appears to be unknown. I pause and notify the class that I will be meeting with them individually during class, knowing that beginning a writing project is sometimes the most difficult part of the process. I begin to conference one-on-one with students: show me your writing process; let’s brainstorm together; where are you struggling; let’s find sources together.
Generative AI tools were explicitly refused in the creation of this blog.
References
Banville, M. & Sugg, J. (2021). “Dataveillance” in the classroom: Advocating for transparency and accountability in college classrooms. Proceedings of the 39th ACM International Conference on Design of Communication, USA, 9–19. https://doi.org/10.1145/3472714.3473617
Banville, M. C. (2023). Am I who I say I am? The illusion of choice: Biometric identification in healthcare (Publication No. 30603350). [Doctoral dissertation, East Carolina University]. ProQuest Dissertations & Theses Global.
Calderwood, S. J. (2025). Surveillance digitisation, performativity, and teacher-student relationships in a blended learning setting. Postdigital Science and Education. https://doi.org/10.1007/s42438-024-00537-6
California State University (CSU). (2026). CSU: AI Commons. https://genai.calstate.edu/
Crooks, R. (2022, October 25). Getting over privacy: Surveillance studies and edtech. Connected Learning Alliance. https://clalliance.org/blog/getting-over-privacy-surveillance-studies-and-edtech/
Edwards, D. (2021). Deep circulation. In E. Beck & L. Hutchinson Campos (Eds.), Privacy Matters: Conversations about surveillance within and beyond the classroom, (pp. 75–92). University Press of Colorado.
Haggerty, K. D., & Ericson, R. V. (2000). The surveillant assemblage. The British Journal of Sociology, 51(4), 605–622. https://doi.org/10.1080/00071310020015280
Neumann, M. D., & Gerstl-Pepin, C. (2025). Faculty responses to generative AI: The shifting landscape of higher education. IEEE International Symposium on Technology and Society (ISTAS), USA, 1–7, https://doi.org/10.1109/ISTAS65609.2025.11269643
Salvo, M. J., & Sherrill, J. T. (2025). Artificial infrastructures. The WAC Clearinghouse. https://doi.org/10.37514/PRA-B.2025.2654
