Open Access
News & Announcements
When AI starts marking assignments and platforms draw learning curves in real time, can a single final exam still decide the outcome? A guide to how digital learning assessment is reshaping teaching and learning
30 October 2026
Innovations in Pedagogy and Technology
Selected Reading Guide
eISSN 3093-7175 (O) | Publisher: NIEP (Nature and Information Engineering Publishing) | Editor-in-Chief: Prof. Arumugam Raman, Universiti Utara Malaysia
Quarterly from 2026 (published in Jan/Apr/Jul/Oct) | Website: https://journals.niepublish.com/ipt
Journal scope: Focused on innovation in pedagogy and educational technology, covering interdisciplinary frontiers such as pedagogy, educational technology, curriculum and instruction, digital learning, educational assessment, and AI applications in education
Volume 02 | Issue 04 | October 2026
When AI starts marking assignments and platforms draw learning curves in real time, can a single final exam still decide the outcome? A guide to how digital learning assessment is reshaping teaching and learning
In the week before final exams, students stay up all night memorizing only to forget, while teachers grade papers until dawn—this old script of “one exam decides all” is all too familiar to us. Now, however, learning management platforms quietly record every log-in, the time spent on each question, and every contribution to a group discussion. Can the data scattered behind these screens be turned into smarter assessment? Learning analytics, automated grading, and early-warning systems—terms that once sounded like science fiction—are now taking hold in pilots at a number of universities and secondary schools. The papers selected in this issue of Innovations in Pedagogy and Technology fit together like a puzzle: how does technology move “assessment” from an end point into the learning process itself? And what unavoidable pitfalls does it leave behind?
01 Why is “one final exam paper” becoming ever less adequate?
A single exam paper can capture only a “mug shot” of learning—a static snapshot of one day, one two-hour sitting—that fails to show how a child climbs through an entire semester or where they get stuck. A study in this issue shows that after a district adopted learning dashboards, teachers could see students’ continuous behaviour on the platform: at which minute of a video their attention wandered, which step of an assignment they repeatedly got wrong, and whether they were active or silent in the discussion forum. Based on this, the system flags students showing warning signs in yellow and those at high risk of falling behind in red, producing an early-warning list even before midterms. Where class teachers once relied on intuition to spot struggling students and often realised only once grades were released, process-based assessment is no longer just a slogan in a lesson plan but a “learning profile” updated daily.
02 Can AI marking really take over essay grading from teachers?
When it comes to reducing workload, what teachers most look forward to is having assignment marking taken off their hands. An essay notebook for a class of around fifty, marked word by word, takes several nights to get through. The paper at the centre of this issue examines pilots of automated essay scoring and intelligent feedback systems: after AI marking was introduced, feedback on objective questions and first drafts of essays returned almost in real time, so students learned the same day where their logic broke down or their evidence was thin. Teachers’ marking time fell on average by about forty per cent, freeing them to devote the saved effort to one-to-one face-to-face coaching. Researchers caution, however, that while AI excels at marking question types with clear rules, its scores remain somewhat “formulaic” when it comes to creative writing or speculative expression, so the final grade still has to fall back on the teacher.
03 As learning profiles grow ever more accurate, who safeguards privacy and fairness?
It all sounds appealing, but would you entrust your child’s every daily mouse-click to an algorithm? The finer the learning profile, the more sensitive the data collected: log-in times, browsing trails, and even—on some platforms—students’ facial expressions and gaze captured through the camera. Case studies warn that warning models may quietly carry bias—if most of the training data comes from better-resourced schools, the algorithm is more likely to misjudge a child who studies on a phone late at night as having “low engagement.” The data divide has merely changed its clothing: students in remote areas have fewer online resources and sparser behavioural data, making them more likely to be “misread” by the algorithm. Some pilot schools have therefore set a rule: raw behavioural data is used only for end-of-term analysis, never entered into progression files, and warning results may trigger intervention only after manual review by a teacher.
04 As teachers move from “judge” to “coach,” does their data literacy keep up?
Once assessment is moved into the process, the first role to change is that of the teacher. A teacher used to be like a judge, delivering a score at the end of term and closing the case; now the role is more like a coach, watching a dynamic curve and calling time-outs or adjusting tactics at any moment. A study in this issue on teacher training finds that many front-line teachers, faced with heat maps and completion-rate curves on a dashboard, are daunted at first—the data is plainly there, but how to read it, at what point something counts as “abnormal,” and which step to intervene on once an anomaly is spotted—no one teaches them hand in hand. The study followed a group of teachers who took part in data-assessment workshops and found that after four to six rounds of hands-on practice, the frequency with which teachers used data for personalised intervention rose markedly, and the rate at which students fell behind also declined. The conclusion is simple: the tools are not lacking—what is missing is a “translator.”
05 From pilot to normal practice, how far still to go?
If the benefits are so obvious, why has it not yet been rolled out across entire schools? Because every step of implementation has to be costed. A review paper in this issue, after surveying more than a dozen digital assessment pilots worldwide, found that the cases that truly closed the loop had generally got three things right: first, data could flow between platforms; second, a dedicated technical support team was available on call, rather than a single orientation session at the start of term and nothing thereafter; third, the school had early on drawn clear ethical boundaries for data use. One detail in the report is thought-provoking: some schools spent large sums installing a system, yet required teachers to fill in half the data by hand, turning the dashboard into a new burden instead. The more technology you add, the more restraint it takes.
Conclusion: why did this issue choose to publish this set of studies?
From a single final exam paper to a learning profile updated daily, the hand of assessment is changing from “passing judgment” to “providing escort.” It ensures that late-blooming students are not buried by one off-day exam, and that early signals of a student falling behind are heard sooner. Yet the cleverer the technology, the more it tests one’s sense of proportion: data must see people, not trap them; AI must lift teachers up, not replace the warm, human parts of education.
To treat this reading guide merely as a commentary on the hot topic of digital learning assessment would be to underestimate its weight. This journal focuses on innovation in pedagogy and educational technology, covering interdisciplinary frontiers such as pedagogy, educational technology, curriculum and instruction, digital learning, educational assessment, and AI applications in education. Digital formative assessment sits precisely at the intersection of the two broad directions of “AI applications in education” and “educational assessment.” This issue selects these papers precisely because they come with real pilot data, front-line scenarios, and pathways to implementation—highly consistent with the journal’s editorial standard of “pursuing new theories, new methods, and practical cases in parallel.”
Publishing this set of studies carries multiple kinds of value. For the discipline, it pulls assessment back from the end point of “one exam decides all” to the whole process of teaching and learning, offering a panoramic perspective from technology and pedagogy through to ethical governance. For practice, it confronts real front-line problems such as automated marking, learning early-warning, the data divide, and teacher data literacy. For readers, it explains professional issues in plain language, so that administrators, researchers, and front-line teachers alike can understand and apply them.
For the journal’s development, selecting this piece is a vivid footnote to its positioning, conveying to scholars worldwide the orientation of “embracing technology, yet asking how technology serves teaching and learning and safeguards fairness and humanity,” and attracting high-quality submissions in educational assessment, learning analytics, and AI ethics in education. It manifests the journal’s distinctiveness and innovativeness—at a time when narratives of the technology dividend prevail, it does not shy away from contentious issues such as algorithmic bias, privacy boundaries, and the data divide, building its identity through a perspective of “technological rationality plus the humanities of education.” It further relies on open-access platforms and diverse global education contexts to extend its international readership and authorship and broaden its editorial-board and reviewer resources, injecting scholarly attention into the journal now that it has moved to a quarterly schedule.
Continued deep cultivation around the signature theme of “technology empowering the transformation of assessment” will help the journal accumulate a series of outputs, laying a foundation for planning special issues, organising online seminars, and raising citations and visibility. The unique value of this article lies in using digital learning assessment as an entry point to interpret the journal’s core thesis: the ultimate goal of innovation in educational technology is to serve teaching and learning, not to simplify assessment or alienate the work of educating people.
References
1. UNESCO. (2024). Global Framework on Digital Learning Assessment.
2. World Economic Forum (WEF). (2025). White Paper on Learning Analytics and the Ethics of Educational Data.
3. International Council for Open and Distance Education. (2025). AI-Enabled Formative Assessment Practice Brief.
4. OECD. (2024). Students, Machines and Learning: Evidence from Digital Assessment Pilots.
About the Journal
This journal is an open-access international English-language academic journal published by the Nature and Information Engineering Publishing, Malaysia. International Standard Serial Number: 3093-7175 (O); Editor-in-Chief: Prof. Arumugam Raman of Universiti Utara Malaysia; the editorial board brings together outstanding scholars from around the world. It was a semi-annual publication in 2025 and became a quarterly from 2026.
The journal mainly publishes research outputs in the fields of innovation in pedagogy and educational technology, focusing on how new technologies empower educational innovation, and covers interdisciplinary frontiers such as pedagogy, educational technology, curriculum and instruction, digital learning, educational assessment, and AI applications in education, accepting new theories, new methods, and practical cases.
Editorial Team page: https://journals.niepublish.com/ipt/about/editorialTeam
Journal website: https://journals.niepublish.com/ipt
Journal email: contact@zycentre.com