OTE Reconstructing the Assessment System of the Undergraduate Course “Plant Physiology” in the AI Era: A Framework for Authentic and AI-Resilient Task Design
Open accessAbstract
The proliferation of generative artificial intelligence (GenAI) has fundamentally challenged the validity of traditional assessments in higher education. In data-rich, laboratory-based disciplines such as undergraduate plant physiology, conventional product-heavy assessments—like essay writing and closed-book examinations—often fail to capture underlying scientific reasoning and are highly susceptible to GenAI simulation. Drawing upon frameworks of constructive alignment, evaluative judgment, and authentic assessment, this paper argues for a systemic shift from misconduct management to “AI-resilient” structural task design. We propose replacing the traditional assessment paradigm with an integrated, seven-part ecology of multimodal tasks. This reconstructed system distributes evidence of learning across the semester through in-class rapid reasoning exercises, process-oriented laboratory notebooks, collaborative inquiry projects, interactive oral defenses (mini-vivas), and reflective AI-use portfolios. By prioritizing visible reasoning, structural verification of knowledge, and critical AI literacy over unsupervised textual output, this framework ensures that students develop and demonstrate authentic disciplinary competence. Ultimately, this approach not only safeguards academic integrity but also enhances the educational validity of plant science curricula, offering a scalable template for assessment reform across the biosciences.
AI snapshot
Written from the paper’s full text.
TL;DR
This conceptual paper argues that generative AI has shifted the assessment problem in undergraduate plant physiology from misconduct management to validity, and proposes replacing product-heavy assessments (essays, closed-book exams, polished lab reports) with a seven-part, semester-long ecology of authentic, "AI-resilient" tasks. The framework distributes evidence of learning across in-class reasoning, process documentation, oral defenses, collaborative inquiry, and reflective AI-use portfolios, but is not empirically tested.
Key findings
- The paper proposes a cumulative seven-part assessment structure replacing practical reports, a term paper, and a final examination: a week-1 diagnostic concept + AI-use baseline (hurdle, unweighted); weekly in-class rapid reasoning tasks (10%); laboratory notebook + raw-data dossier (15%); two mini-vivas after laboratory blocks (15%); group inquiry project (20%); individual integrative oral exam (25%); reflective portfolio on feedback and AI use (15%).
- The rationale is that AI-resilience comes from structure rather than prohibition: tasks are time-limited and supervised, grounded in localized lab activity, or require live explanation, so "generic AI substitution" is ineffective.
- Six illustrative task designs are offered for plant physiology, including photosynthesis limitation diagnosis with a falsification follow-up, a two-minute water-relations emergency brief, a mineral nutrition diagnostic memo plus viva, a hormone-signaling experimental design studio, an abiotic stress case conference using imperfect data, and an AI critique/verification memo containing subtle flaws such as causal leaps and fabricated references.
- The paper claims three strengths for the reconstruction: improved validity (alignment with mechanistic reasoning and experimental thinking), improved integrity (structural verification of understanding), and enhanced educational relevance (critical AI literacy).
Methods
- Design: conceptual/theoretical synthesis drawing on constructive alignment (Biggs, 1996), assessment for learning and feedforward (Boud & Falchikov, 2006; Carless, 2007), evaluative judgment (Tai et al., 2018; Bearman et al., 2024), authentic assessment (Gulikers et al., 2004; Fawns et al., 2025; McArthur, 2023), and AI literacy (Lintner, 2024).
- Data or sample: none; the paper presents a proposed framework, a component/weighting table, and worked task examples rather than collected data.
- Main analysis: argumentative development of design principles (visible reasoning, distributed evidence, collaboration with individual accountability, local/novel/messy evidence, assessable AI critique, structural inclusion) and their translation into course components, rubrics, feedback, moderation, and workload considerations.
Limitations
- Stated by the paper: authenticity alone is insufficient, since even authentic-seeming assessments can be outsourced (Ellis et al., 2020), and the design "works only as a holistic system" where distributed evidence, process documentation, oral defense, and AI literacy reinforce one another; the authors call for future empirical studies on effects on student learning, anxiety, misconduct patterns, and AI literacy.
- Evident from the text: no empirical validation, pilot data, or student outcomes are reported for the proposed weighting scheme or task examples; the framework is developed for a single course and discipline (undergraduate plant physiology), and the paper itself notes that broad conceptual frameworks must be translated into discipline-specific solutions, so scalability to other bioscience courses is asserted rather than demonstrated.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
15 of 24 cited sources checked · 9 supported · 0 flagged · accuracy not yet measured
- Premisecitation
Pairing group inquiry with individual oral defenses preserves authenticity while protecting validity.
“Pairing group inquiry with individual oral defenses preserves authenticity while protecting validity (Davey et al., 2025;Ward et al., 2024).”
Quote in the paperSubstantial claim2 of 2 sources checked- cites Davey et al., 2025 source
The Discussion reports that students studied in groups before an individualized oral assessment, which distinguished collaborative learning from collusion. Theme 1 (Authenticity) says the oral format let assessors probe individual understanding, and Theme 2 (Academic integrity) reports resistance to misconduct. The paper does not describe pairing a group inquiry task with individual oral defenses; its group activity was assessment preparation. (LLM, conf=0.85)
Experimental — accuracy not yet measured - cites Ward et al., 2024 source
The Computing case study (page unavailable) describes teams conducting background research and discussing their recommendations in a group interactive oral, with students marked individually. The Discussion calls these assessments authentic and academically rigorous and says individual grading addressed fairness concerns. It does not describe separate individual oral defenses, so the claimed pairing is more specific than the source supports. (LLM, conf=0.84)
Experimental — accuracy not yet measured
- Premisecitation
Tasks requiring genuine collaboration offer greater resistance to AI manipulation.
“Tasks requiring genuine collaboration offer greater resistance to AI manipulation (Bannister et al., 2025).”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Bannister et al., 2025 source
- Premisecitation
Large language models generate plausible essays, summaries, laboratory write-ups, and examination responses with limited student engagement.
“As large language models generate plausible essays, summaries, laboratory write-ups, and examination responses with limited student engagement (Ahangama, 2026)”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Ahangama, 2026 source
Theme 1–Values & Dispositions (page unavailable) discusses students submitting AI-generated content as their own work, and Theme 7–Assessment Methods describes staged essays intended to foster authentic engagement. These support the general concern that AI can substitute for student work, but our copy does not substantiate the specific list of summaries, laboratory write-ups, and examination responses, or establish that LLMs produce all of them plausibly with limited student engagement. (LLM, conf=0.78)
Experimental — accuracy not yet measured
- Premisecitation
Evaluative judgment requires students to recognize quality in AI-mediated processes.
“Evaluative judgment requires students to recognize quality in AI-mediated processes (Bearman et al., 2024).”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Bearman et al., 2024 source
- Premisecitation
Heavy dependence on a single final task is pedagogically weak.
“Heavy dependence on a single final task is pedagogically weak (French et al., 2024).”
Quote in the paperSubstantial claim1 of 1 sources checked- cites French et al., 2024 source
- Premisecitation
Studies show no significant performance differences across demographics when students are carefully scaffolded.
“studies show no significant performance differences across demographics when students are carefully scaffolded (Davey et al., 2025).”
Quote in the paperNot a substantial claim1 of 1 sources checked- cites Davey et al., 2025 source
- Premisecitation
Plant-science education research demonstrates that inquiry-rich, research-infused, and communication-centered approaches significantly improve knowledge, reasoning, writing, and disciplinary interest.
“Plant-science education research demonstrates that inquiry-rich, research-infused, and communication-centered approaches significantly improve knowledge, reasoning, writing, and disciplinary interest (Beckmann et al., 2015;Hiatt et al., 2021;Ward et al., 2014).”
Quote in the paperSubstantial claim2 of 3 sources checked- cites Beckmann et al., 2015 source
The Results (page unavailable) report improved theoretical learning, student accounts of stronger scientific reasoning, and increased progression into plant-science Honors, supporting much of the claim. The Methods describe writing instruction and written assessment, but the paper does not report a measured improvement in writing. Its evaluation relied largely on qualitative feedback, so the claim that these approaches significantly improved all four outcomes overstates the evidence. (LLM, conf=0.88)
Experimental — accuracy not yet measured - cites Hiatt et al., 2021 source
- cites Ward et al., 2014 source
The Results sections on Content Knowledge and Scientific Practice report significant gains in botanical knowledge, statistical reasoning, and scientific writing. The Attitudinal Shifts section reports greater awareness of plants and interest in research, but intent to take additional plant courses—the paper’s measure of botanical interest—was unchanged (p = 0.76). (LLM, conf=0.90)
Experimental — accuracy not yet measured
- Premisecitation
Research highlights the difficulty of distinguishing AI-generated work from student writing.
“Research highlights the difficulty of distinguishing AI-generated work from student writing (Farazouli et al., 2024)”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Farazouli et al., 2024 source
- Premisecitation
The proliferation of generative artificial intelligence (GenAI) has fundamentally challenged the validity of traditional assessments in higher education.
“The proliferation of generative artificial intelligence (GenAI) has fundamentally challenged the validity of traditional assessments in higher education.”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Ahangama, 2026 source
- Interpretationreasoning
Given its empirical and interpretive core, plant physiology is well-suited for process-oriented and multimodal assessment reform.
“Given its empirical and interpretive core, plant physiology is well-suited for process-oriented and multimodal assessment reform.”
Quote in the paperNot a substantial claim - Premisecitation
Authenticity or policy alone cannot guarantee integrity.
“Because authenticity or policy alone cannot guarantee integrity (Ellis et al., 2020;Fawns et al., 2025)”
Quote in the paperSubstantial claim0 of 2 sources checked- cites Ellis et al., 2020
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- cites Fawns et al., 2025 source
- Interpretationreasoning
The proposed approach safeguards academic integrity.
“Ultimately, this approach not only safeguards academic integrity but also enhances the educational validity of plant science curricula”
Quote in the paperSubstantial claim - Interpretationreasoning
Conventional assessment formats no longer provide robust evidence of student knowledge, capability, and justification.
“The deeper concern is that conventional assessment formats no longer provide robust evidence of student knowledge, capability, and justification.”
Quote in the paperSubstantial claim - Interpretationreasoning
GenAI acts as a catalyst exposing the mismatch between traditional product-heavy assessments and the discipline's inquiry-based goals.
“GenAI acts as a catalyst exposing the mismatch between traditional product-heavy assessments and the discipline's inquiry-based goals.”
Quote in the paperNot a substantial claim - Premisecitation
Oral and authentic assessments are not automatically inclusive.
“Oral and authentic assessments are not automatically inclusive (Fawns et al., 2025;McArthur, 2023).”
Quote in the paperSubstantial claim0 of 2 sources checked- cites Fawns et al., 2025 source
- cites McArthur, 2023
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- May combine two claims. 6
- Classified as probably not a substantial claim. 25
May combine two claims.
PremisecitationEvaluating unseen and situated evidence diminishes the temptation to plagiarize and prevents cognitive offloading to AI.
“This diminishes the temptation to plagiarize and prevents cognitive offloading to AI (Peters et al., 2025).”
Quote in the paperSubstantial claim1 of 1 sources checked- cites Peters et al., 2025 source
The 'Authentic assessment tasks' section (page unavailable) describes authentic assessment as a deterrent to plagiarism, while noting that it does not always prevent cheating. The 'Cognitive offloading' section (page unavailable) advocates deciding when and how students may use GenAI for offloading, not preventing it. The copy therefore supports the plagiarism-deterrence part, but not the claim that evaluating unseen and situated evidence prevents cognitive offloading to AI. (LLM, conf=0.88)
Experimental — accuracy not yet measured
Probably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationBalancing detailed guidance with student autonomy helps avoid inadvertently facilitating GenAI prompt formulation.
“Furthermore, balancing detailed guidance with student autonomy helps avoid inadvertently facilitating GenAI prompt formulation (Bannister et al., 2025).”
Quote in the paperNot a substantial claim1 of 1 sources checked- cites Bannister et al., 2025 source
Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).
PremisecitationWorkload concerns are valid.
“while workload concerns are valid (Corbin et al., 2025a)”
Quote in the paperNot a substantial claim1 of 1 sources checked- cites Corbin et al., 2025a source
May combine two claims.
PremisecitationEvidence suggests the pedagogical justification for high-stakes final examinations is weak, with significant drawbacks for authenticity, wellbeing, equity, and long-term learning.
“Evidence suggests the pedagogical justification for high-stakes final examinations is weak, with significant drawbacks for authenticity, wellbeing, equity, and long-term learning (French et al., 2024).”
Quote in the paperSubstantial claim1 of 1 sources checked- cites French et al., 2024 source
Probably not a substantial claim: may not pass S2 (checkable), S4 (based).
InterpretationreasoningThe design works only as a holistic system where distributed evidence, process documentation, oral defense, and AI literacy reinforce one another.
“The design works only as a holistic system where distributed evidence, process documentation, oral defense, and AI literacy reinforce one another.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S4 (based).
InterpretationreasoningA reconstructed assessment ecology integrating in-class reasoning, process evidence, oral defense, and reflective AI-use documentation is more faithful to disciplinary practice.
“A reconstructed assessment ecology integrating in-class reasoning, process evidence, oral defense, and reflective AI-use documentation is more faithful to disciplinary practice”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based). May combine two claims.
Interpretationnone givenLaboratory notebook + raw-data dossier is hard to outsource and grounded in actual, localized lab activity.
“Hard to outsource; grounded in actual, localized lab activity.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationThe proposed course aligns with structural redesign recommendations in AI-assessment literature.
“This aligns with structural redesign recommendations in AI-assessment literature (Corbin et al., 2025b).”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites Corbin et al., 2025b source
Probably not a substantial claim: may not pass S2 (checkable).
InterpretationreasoningA reconstructed assessment ecology integrating in-class reasoning, process evidence, oral defense, and reflective AI-use documentation is more defensible against AI misuse.
“A reconstructed assessment ecology integrating in-class reasoning, process evidence, oral defense, and reflective AI-use documentation is more faithful to disciplinary practice, more defensible against AI misuse, and ultimately more educationally valuable.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable).
InterpretationreasoningA reconstructed assessment ecology integrating in-class reasoning, process evidence, oral defense, and reflective AI-use documentation is ultimately more educationally valuable.
“A reconstructed assessment ecology integrating in-class reasoning, process evidence, oral defense, and reflective AI-use documentation is more faithful to disciplinary practice, more defensible against AI misuse, and ultimately more educationally valuable.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based). May combine two claims.
Interpretationnone givenThe following examples demonstrate assessment tasks that are difficult to outsource and reliant on contextual judgment.
“The following examples demonstrate assessment tasks that are difficult to outsource and reliant on contextual judgment.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S4 (based).
Interpretationnone givenThe proposed reconstruction enhances educational relevance by developing critical AI literacy.
“it enhances educational relevance by developing critical AI literacy.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing).
PremisecitationCourse-based undergraduate research experiences shift awareness and interest in plants.
“while course-based undergraduate research experiences shift awareness and interest in plants (Hiatt et al., 2021;Ward et al., 2014)”
Quote in the paperNot a substantial claim0 of 2 sources checkedProbably not a substantial claim: may not pass S2 (checkable), S4 (based).
Interpretationnone givenPlant physiology depends on interpretive and experimental capacities that are difficult to infer from finished text alone.
“Unlike courses where success is demonstrated through polished exposition, plant physiology depends on interpretive and experimental capacities that are difficult to infer from finished text alone.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).
InterpretationreasoningRedesigns can redistribute labor productively.
“redesigns can redistribute labor productively.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based).
Interpretationnone givenFollow-up questioning in mini-vivas exposes shallow or borrowed knowledge.
“Follow-up questioning exposes shallow or borrowed knowledge.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based).
InterpretationreasoningThe proposed reconstruction improves validity by aligning assessment with mechanistic reasoning and experimental thinking.
“it improves validity by aligning assessment with mechanistic reasoning and experimental thinking”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S4 (based).
InterpretationreasoningGenAI magnifies these weaknesses by easily simulating competence without requiring the underlying scientific process.
“GenAI magnifies these weaknesses by easily simulating competence without requiring the underlying scientific process.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based).
Interpretationnone givenLive explanation in individual integrative oral exam makes generic AI substitution ineffective.
“Live explanation makes generic AI substitution ineffective.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based). May combine two claims.
Interpretationnone givenIn data-rich, laboratory-based disciplines such as undergraduate plant physiology, conventional product-heavy assessments (essay writing and closed-book examinations) often fail to capture underlying scientific reasoning and are highly susceptible to GenAI simulation.
“In data-rich, laboratory-based disciplines such as undergraduate plant physiology, conventional product-heavy assessments-like essay writing and closed-book examinations-often fail to capture underlying scientific reasoning and are highly susceptible to GenAI simulation.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based).
Interpretationnone givenHistorically, science assessments have over-weighted end-point products: polished lab reports, literature reviews, and closed-book examinations.
“Historically, science assessments have over-weighted end-point products: polished lab reports, literature reviews, and closed-book examinations.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S4 (based).
InterpretationreasoningThe proposed framework ensures that students develop and demonstrate authentic disciplinary competence by prioritizing visible reasoning, structural verification of knowledge, and critical AI literacy over unsupervised textual output.
“By prioritizing visible reasoning, structural verification of knowledge, and critical AI literacy over unsupervised textual output, this framework ensures that students develop and demonstrate authentic disciplinary competence.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S4 (based).
Interpretationnone givenThe GenAI era has exposed the fragility of systems that over-rely on unsupervised, polished, end-point textual production.
“It has exposed the fragility of systems that over-rely on unsupervised, polished, end-point textual production.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S4 (based).
Interpretationnone givenThe proposed reconstruction improves integrity by structurally verifying understanding.
“it improves integrity by structurally verifying understanding”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable).
InterpretationreasoningThe proposed approach enhances the educational validity of plant science curricula.
“Ultimately, this approach not only safeguards academic integrity but also enhances the educational validity of plant science curricula”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S2 (checkable), S4 (based).
InterpretationreasoningThese formats create distortions by rewarding retrospective neatness or short-term retrieval over iterative inquiry.
“These formats create distortions by rewarding retrospective neatness or short-term retrieval over iterative inquiry.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
InterpretationcitationShort in-class tasks replace the grading of long take-home essays, and notebook assessments can be streamlined through selective review, shifting faculty effort.
“Short in-class tasks replace the grading of long take-home essays, and notebook assessments can be streamlined through selective review (Fleming, 2023), shifting faculty effort from”
Quote in the paperNot a substantial claim0 of 1 sources checked- cites Fleming, 2023 source
Checked before — its earlier verdict is held back until a new check runs.
Cites
The publisher hasn’t deposited this paper’s reference list.
- Authors
- Peng, J.
- Affiliations
- Guizhou University (CN)
- Corresponding author
- Peng, Jiantao
- Venue
- World Journal of Educational Research
- Publisher
- Scholink Co, Ltd.
- Published
- 2026-04-24
- Volume
- 13
- Issue
- 2
- Pages
- p127
- ISSN
- 2333-5998 (ISSN-L)
- 2375-9771
- OpenAlex
- W7155513664
- Semantic Scholar
- 381516c4f1a09e2e8c78ed861074e5ad7cfcdaa9
- Cited by
- 0
- References
- 0
- FWCI
- 0.00
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Locations
- World Journal of Educational Research (publishedVersion, cc-by)
- Topics
- Social Sciences › Education › Student Assessment and Feedback
- Social Sciences › Education › Educational Assessment and Pedagogy
- Social Sciences › Safety Research › Academic integrity and plagiarism
- Keywords
- plant physiology education, assessment reform, generative AI, authentic assessment, constructive alignment, evaluative judgments, academic integrity, AI literacy, laboratory notebooks, oral examination, reflective portfolio, critical thinking, biology education
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.
Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-29.
Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-29.
Cited sources are checked by gpt-6-sol.
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
