The DOI matches its Crossref record; 5 of the 5 references it cites were found in a registry or an index; none of the 33 sources its claims cite is checked for support yet.
Reimagining the Artificial Intelligence Assessment Scale: A refined framework for educational assessment
Open accessAbstract
Higher education institutions, educators, and students continue to grapple with the wide-ranging implications of Artificial Intelligence (AI) and Generative Artificial Intelligence (GenAI). The ability of advanced GenAI models to complete educational assessments continues to be one of the most pressing issues for academia. In early 2024, we published the AI Assessment Scale (AIAS), which describes a practical framework for addressing GenAI use in educational assessment. The AIAS was well received and has been implemented in hundreds of institutions worldwide and translated into 30 languages. Building on our experience using the AIAS and drawing on feedback and critiques received worldwide, in this paper, we provide an updated version of the AIAS. The updated AIAS encompasses several important developments, including clarifying the theoretical basis of the scale through social constructivist principles, developing the AIAS to function as an assessment redesign framework to enhance validity, and adjusting the visual assets of the AIAS to function inclusively and non-hierarchically. Furthermore, in anticipation of future developments in the field of AI and the increasing focus on preparing students to function in an AI-augmented world, we introduce an additional level, ‘AI exploration’. Through this paper we also seek to clarify both the intended and unintended applications of the AIAS. We provide implementation guidance through practical examples and explain how the AIAS functions as both a communication tool and a framework for task redesign that strengthens validity.
Overview
CoChat Summary
## TL;DR
This paper revises the Artificial Intelligence Assessment Scale (AIAS) into a five-level framework for assessment redesign in an AI-augmented educational environment. The update adds an “AI exploration” level, replaces the traffic-light visual scheme with a neutral palette/circular representation, and emphasizes validity, transparency, accessibility, and social constructivist principles.
## Key findings
- The original AIAS had been implemented in hundreds of institutions worldwide and translated into 30 languages; the paper says the revised version builds on feedback and critiques from those uses.
- The revised AIAS retains five levels but updates them to: Level 1 No AI, Level 2 AI-Assisted Planning, Level 3 AI-Assisted Task Completion, Level 4 Full AI, and Level 5 AI Exploration.
- The authors state that the revised framework shifts away from “security theatre” and instead requires substantive structural assessment redesign, including changes to task mechanics, checkpoints, evidence of process, and marking criteria.
- The paper says the updated framework better reflects current GenAI capabilities, including multimodal content generation across text, images, audio, and video, and clarifies that no level is inherently better than another.
## Methods
- This is a conceptual/framework paper based on the authors’ experience using the AIAS, worldwide feedback and critiques, and implementation experiences across contexts.
- The paper draws on social constructivist principles and Vygotsky’s framework to justify the AIAS as an assessment redesign tool.
- The main analysis is iterative refinement of the scale: the authors compare the original and updated AIAS, justify changes to each level, and provide example vignettes of how assessments might look at each level.
## Limitations
- The paper explicitly states that further empirical validation is crucial and calls for future research on effectiveness across different educational contexts, as well as qualitative and quantitative studies of student/educator experiences, learning outcomes, and assessment validity.
- The authors acknowledge that the framework’s equity solutions are imperfect, especially around access to GenAI tools at Levels 4 and 5.
- The text also makes clear that the paper is not an empirical evaluation of the revised AIAS itself; it is a revision and rationale paper, so its claims about effectiveness are grounded in prior pilot evidence and argument rather than new data (evident from the text).
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
0 Valid · 0 Not valid · 26 Not checked · 65 Own results and interpretations
In the paper
“spend more time 'thinking' or 'reasoning', enabling them to achieve top percentile scores in mathematical and coding competitions and challenges (OpenAI, 2024).”
What the sources say
- Not matchedcites OpenAI, 2024
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Not matched
A conclusion the paper draws from its own results — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This means that assessment tasks that were initially thought to be more resistant to GenAI tools are also vulnerable to completion by these tools.”
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
None of the sources cited for this claim have been checked yet.
In the paper
“GenAI text detectors are commonly used as adversarial tools to identify academic integrity violations by students, but these tools are unreliable (Elkhatat et al., 2023; Perkins, Roe, et al., 2024; Sadasivan et al., 2023; Weber-Wulff et al., 2023).”
What the sources say
- Not matchedcites Elkhatat et al., 2023
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not checked yetcites Perkins, Roe, et al., 2024 source
- Not matchedcites Sadasivan et al., 2023
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Not matched
None of the sources cited for this claim could be matched to a paper in Crossref, so there is nothing to check it against.
In the paper
“Since the initial release of ChatGPT, GenAI technology has become much more capable across an ever-broadening range of writing, mathematical, and scientific reasoning tasks.”
What the sources say
- Not matchedcites Head & Willis, 2024
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not matchedcites Newton et al., 2024
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Not matchedcites Shahriar et al., 2024
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Not matched
In the paper
“GenAI models can now handle multimodal content effectively (Shahriar et al., 2024)”
What the sources say
- Not matchedcites Shahriar et al., 2024
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Not matched
13 more claims
Create a free account to see all 91 claims and to verify the 2 references still open.
- Probably not a substantial claim. 68
- It may state two separate claims as one. 28
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ Critique of the AIAS
“| AI Assessment Scale | Two-Lane |”
§ Critique of the AIAS
“| Lane 1 - Secure Assessments | Level 1 - No AI |”
Citations
Cites
The publisher hasn’t deposited this paper’s reference list.
Cited by
The EAP‐AIAS: Adapting the AI Assessment Scale for English for Academic Purposes
Roe, J., Perkins, M., & Tregubova, Y. (2026) · TESOL Journal · doi:10.1002/tesj.70122 · Cited by 38
Self-citationSu, F., O'Dea, M., Wood, M., & Seidl, E. (2026) · Perspectives Policy and Practice in Higher Education · doi:10.1080/13603108.2025.2601741 · Cited by 12
Applying the AI assessment scale in writing and translation for EFL
Roe, J., Perkins, M., & Furze, L. (2025) · Artificial Intelligence in Education · doi:10.1108/aiie-02-2025-0034 · Cited by 21
Self-citationMoore, J., Frangieh, J., Capello, A., Montejo, L., Lukkahatai, N., Aryal, S., Zhang, J., & Mudd, S. (2026) · Nursing Outlook · doi:10.1016/j.outlook.2026.102856 · Cited by 2
Fawns, T., Bearman, M., Corbin, T., Henderson, M., McLean, J., Matthews, K., Oberg, G., Liang, Y., & Walton, J. (2026) · Higher Education · doi:10.1007/s10734-026-01730-1 · Cited by 2
Zeivots, S., Wright, S., Harris, L., Cram, A., Huber, E., Raduescu, C., White, A. W., & Brodzeli, A. (2026) · Studies in Higher Education · doi:10.1080/03075079.2026.2619909 · Cited by 3
Ateşkan, A. (2026) · Assessment & Evaluation in Higher Education · doi:10.1080/02602938.2026.2678375 · Cited by 1
Ruby, M. F. (2026) · Research Square · doi:10.21203/rs.3.rs-8800271/v1
TRANSFORMATION OF LEARNING TASKS USING GENERATIVE AI
Kovalenko, I. (2026) · OPEN EDUCATIONAL E-ENVIRONMENT OF MODERN UNIVERSITY · doi:10.28925/2414-0325.2026.204 · Cited by 2
‘A bit of chaos and madness’: the AI assessment scale and the work of assessment reform
Perkins, M., Postma, D., Roe, J., Sisay, S., & Holdcroft, C. (2026) · Assessment & Evaluation in Higher Education · doi:10.1080/02602938.2026.2724300 · Cited by 1
“When it is issued without these supports, it risks becoming another discursive response that leaves the underlying validity problem untouched (Corbin, Dawson, & Liu, 2025; Perkins et al., 2025a).”
InfluentialSelf-citationAlvarado, A., & Portillo, J. R. (2026) · IEEE Global Engineering Education Conference (EDUCON) · doi:10.1109/educon67543.2026.11574397
From manual to autonomous: a literature review of GenAI integration levels for design education
Mao, L., Qi, M., Ho, J. C. F., Bruyns, G., Lee, K.-P., & Zhou, Z. (2026) · Interactive Learning Environments · doi:10.1080/10494820.2026.2680597
Enabling Student Ownership of LLM-Use through Team Contracts
Westerman, F., & Carruthers, S. (2026) · ACM Conference on Innovation and Technology in Computer Science Education (ITiCSE) · doi:10.1145/3803401.3811958
Wong, A. K. L., Tsang, K. C. K., Lai, W., Lin, S., & Chan, L. L. (2026) · Education Sciences · doi:10.3390/educsci16081193
Dumbrăveanu, R., Grosseck, G., Maliţa, L., & BALMUȘ, O. (2026) · International Conference on e-Education, e-Business, e-Management and e-Learning · doi:10.1109/ic4e69618.2026.11631500
“By adopting flexible strategies, such as a 'traffic light' system for assessments [17] and prioritizing ethical AI literacy over punitive measures, institutions can ensure that academic integrity is maintained without stifling the essential development of students' digital competencies.”
Vu, H. T. T., Phan, H. T., Tran, M. C. T., Le, V. A., Ho, N. V., Doan, M. H. T., Nguyen, T. P. T., & Pham, T. H. (2026) · Social Sciences & Humanities Open · doi:10.1016/j.ssaho.2026.103462 · Cited by 1
Jose, L. S., & Cabual, R. A. (2026) · International Journal of Learning Teaching and Educational Research · doi:10.26803/ijlter.25.8.37
Structural Assessment Redesign for AI-Allowed Online Learning through Ethical AwarenessPreprint
Doğan, S. (2026) · Research Square · doi:10.21203/rs.3.rs-10829938/v1
10 insights for harnessing artificial intelligence in scientific research: supervisor perspectives
Mitra, N. K., Vishnumukkala, T., & Chitra, E. (2026) · MedEdPublish · doi:10.12688/mep.21799.1
Bartz, A., Glaser, N., Bayer, O., Hallsson, L. R., Jaeger, V. K., Krumkamp, R., Minnerup, H., Pickford, R., Olsen, K. S., Wülk, N., et al. (2026) · Research Square · doi:10.21203/rs.3.rs-10974790/v1
von Maltitz, M. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-9089-5.ch012
Hochhauser, M., & Akselrud, R. (2026) · British Journal of Occupational Therapy · doi:10.1177/03080226261484046
Sharma, S., & Choudhary, A. (2026) · Assessment in Education Principles Policy and Practice · doi:10.1080/0969594x.2026.2736629
Principles without provisions? Unpacking assessment in higher education AI policies
Egloffstein, M., & Heil, J. (2026) · Assessment in Education Principles Policy and Practice · doi:10.1080/0969594x.2026.2740061
Педагогічні ризики когнітивного делегування генеративному штучному інтелекту у вищій освіті
Mudryk, D. P. (2026) · Проблеми сучасних трансформацій Серія педагогіка · doi:10.54929/2786-9199-2026-13-03-09
Villanueva, G. (2026) · Assessment & Evaluation in Higher Education · doi:10.1080/02602938.2026.2740764
Plaatjies, B. O., & van Wyk, M. M. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-8332-3.ch002
Metadata
- Authors
- Perkins, M., Jasper, R., & Furze, L.
- Affiliations
- British University Vietnam (VN)
- Deakin University (AU)
- Venue
- Journal of University Teaching and Learning Practice
- Publisher
- Open Access Publishing Association
- Published
- 2025-12-03
- Volume
- 22
- Issue
- 7
- ISSN
- 1449-9789 (ISSN-L)
- OpenAlex
- W4414552639
- Cited by
- 30
- References
- 0
- Citations by year
- 2026: 29 · 2025: 1
- FWCI
- 31.06
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- Locations
- Journal of University Teaching and Learning Practice (publishedVersion, cc-by-nc-sa)
- Topics
- Social Sciences › Education › Artificial Intelligence in Education
- Social Sciences › Education › Academic integrity and plagiarism
- Social Sciences › Education › Student Assessment and Feedback
- Keywords
- generative AI, assessment redesign, scale revision, social constructivism, assessment validity, AI-augmented learning, higher education assessment, inclusive design
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.