Why surveillance-based assessment doesn't work at scale
The default response to AI in education has been to ramp up detection—plagiarism checkers, authorship verification, and monitoring tools. But this approach has fundamental flaws that make it unsustainable at scale. A systematic review of studies from 2020 to 2025 found that AI-driven integrity tools still suffer from bias, false positives, and limited transparency, meaning they can wrongly accuse students or miss sophisticated AI use entirely [1]. Faculty in a multisite study of 36 academics across three universities explicitly rejected 'detection-led policing' as a strategy, arguing it undermines trust and does not address the root problem [4]. The evidence is clear: surveillance creates an arms race that erodes the educational relationship and fails to keep pace with rapidly evolving AI tools.
The proven alternative: assessments that reward process, not just product
This approach scales because it doesn't depend on catching cheaters; it makes cheating irrelevant by design. When assessments require human judgment, personal reflection, or real-time collaboration, AI becomes a partner rather than a replacement. The studies agree that this shift also prepares students for careers where AI will be a standard tool, making the assessment more relevant to their future work [2][5].
What institutions must do to make this work at scale
Scaling process-oriented assessment without surveillance isn't automatic—it requires institutional support. Faculty in the studies identified key enablers: professional development to learn new assessment techniques, guaranteed access to AI tools for all students (to avoid equity gaps), and protected time for redesign work [4][5]. One paper modeled assessment as influenced by ten factors, including market pressures and demand for flexible, student-centered learning, and argued that formative assessment supported by AI and learning analytics can facilitate this shift [2]. The bottom line: institutions must invest in training and infrastructure, not surveillance software. Without that investment, the default will remain policing, which the evidence shows is both ineffective and damaging to trust.
About These Sources
This answer is built on 5 peer-reviewed studies — published from 2023 to 2026, 4 from 2024 or later, 2 in Q1 journals, collectively cited 63 times — selected as the most relevant from 5 studies that passed quality screening, drawn from 44 papers retrieved from a database of over 500 million.
Sources used in this answer
Artificial Intelligence and the Transformation of Academic Integrity in Higher Education: A Systematic Review
A systematic review of studies from 2020-2025 found that AI integrity tools (plagiarism detectors, authorship checkers) still suffer from bias, false positives, and limited transparency, and concluded institutions should focus on ethical digital literacy and assessment redesign rather than surveillance.
AI, Analytics and a New Assessment Model for Universities
A SWOT analysis of current assessment practice identified AI and learning analytics as key factors, arguing that formative assessment supported by these technologies can enable more flexible, student-centered assessment for learning.
Authentic and Creative Assessment in a World with AI
A review of authentic assessment practices provides sample activities where AI is used to design assessments that foster creativity, critical thinking, and reflection, aligning with APA learning goals for psychology majors.
From Policing to Design: A Qualitative Multisite Study of Generative Artificial Intelligence and SDG 4 in Higher Education
A qualitative study of 36 faculty across three universities found staff rejected detection-led policing and favored assessment designs that reward process, critique, and provenance, offering a practical framework aligned with SDG 4 targets.
Redesigning Assessments for AI-Enhanced Learning: A Framework for Educators in the Generative AI Era
Interviews with 61 faculty members identified motivations and challenges for redesigning assessments in the Gen AI era, leading to the 'Against, Avoid, Adopt, Explore' framework that guides educators in choosing appropriate assessment strategies.