AI in Re-imagining Assessment : The Situated Learning Assessment Model (SLAM)

Team name: JGU 5

Use of AI tools : If you used AI, please specify:

• Which tool(s) did you use?
  ChatGPT (OpenAI)

  Claude (Anthropic)

  GitHub Copilot

  Perplexity AI

• For what purpose(s)? (e.g., research, structuring ideas, drafting, data analysis, etc.)
  Structuring and refining ideas

  Literature exploration and preliminary research assistance

  Drafting and editing sections of the proposal

  Improving clarity, coherence, and readability

  Technical debugging and coding assistance

  Benchmarking AI outputs and testing model behavior

  Synthesizing feedback received from contributors and reviewers

  Generating alternative formulations and conceptual frameworks

  Assisting with statistical interpretation and methodological structuring

• How did these tools support your work?
  AI tools primarily functioned as collaborative support systems that accelerated iterative work processes. They helped us organize complex ideas, compare conceptual approaches, refine technical explanations, and improve the accessibility of the proposal’s language. For the technical components, AI-assisted coding tools supported debugging, prototype iteration, and documentation. General-purpose models were also used as comparative baselines in evaluating the feasibility and limitations of our proposed system.

  All substantive decisions, critical analysis, final arguments, evaluation criteria, and conceptual directions were independently reviewed, validated, and selected by the team.

External feedback & contributions :
Prof. Paavni Jain - Assistant Professor, O.P Jindal Global University - Overall Feedback

Prof. Spriha Bhandari- Assistant Professor, O.P Jindal Global University - Overall Feedback

Bruno de Lièvre – Professor, University of Mons – Feedback on conceptual structure, feasibility, and pilot design

Lucie Jacquet-Malo – Coordinator of AI Program, Institut Mines-Télécom – Feedback on scalability, institutional integration, and sustainability

Dr. Christelle Scharff – Academic/Researcher – Discussion on EU AI Act implications, bias mitigation, and human oversight

Nicol D. – Educational Psychology Perspective (PsyEN) – Feedback on the distinction between information and orientation

Hans Zúñiga – Participant/Reviewer – Feedback on transparency, equity, and human-centered guidance

Kathleen Somers – Reviewer – Discussion regarding exploratory orientation mechanisms and profession recommendation logic

Ndeye Sophie Seck – Participant/Reviewer – Feedback on critical thinking, human reflection, and AI dependency concerns

Initial contribution: AI in Re-imagining Assessments

Final contribution: https://drive.google.com/drive/folders/17iQKFJL0fRnPTWho09t9okEinYdMZxZd

And for another part of the submission we have made a website which can be accessed on: https://slamjgu5.netlify.app/

Reflection on the process (important)
The Phase 2 deliberation process significantly strengthened our proposal by forcing us to move beyond a broad critique of AI in education toward a more precise and implementable assessment framework. Feedback from peers, faculty, and other teams helped us refine the core problem as the collapse of output-based evaluation as a reliable proxy for genuine understanding rather than AI use alone. It also pushed us to think more concretely about scalability, institutional safeguards, faculty workload, implementation feasibility, and the distinction between AI literacy and AI dependency. As a result, the final proposal became more balanced, structurally coherent, and institutionally grounded while preserving its central argument that assessment must shift from verifying submissions to verifying demonstrated understanding.

Explain how your contribution evolved during Phase 2:

• feedback that influenced your work?
  Peer discussions, faculty feedback, and exchanges with other teams significantly shaped the evolution of our proposal during Phase 2. A major point repeatedly raised was that our earlier drafts focused too broadly on AI access disparities and plagiarism detection, without sufficiently identifying the deeper structural problem within higher education assessment itself. Feedback also pushed us to think more carefully about scalability, faculty workload, safeguards against subjectivity in viva-based systems, implementation feasibility across institutions with differing capacities, and the distinction between AI literacy and AI dependency. Discussions with other teams also helped us sharpen the idea that generative AI did not create the crisis in education, but rather exposed pre-existing weaknesses in how institutions evaluate understanding.

• changes made as a result?
  As a result of this feedback, we substantially refined both the structure and focus of our proposal. We shifted the core argument away from AI as merely a plagiarism or integrity issue and reframed it as the collapse of output-based assessment as a reliable proxy for genuine understanding. The SLAM framework was also redesigned to become more operational and institutionally grounded. We incorporated phased implementation models, contextual calibration mechanisms, faculty moderation systems, auditability safeguards, and clearer distinctions between format and standard within evaluation. We also clarified that the proposal is not anti-AI, and explicitly incorporated AI literacy as a legitimate educational competency while maintaining that demonstrated cognition must remain central to assessment.

• How this process strengthened the final proposal?
  This process strengthened the proposal by forcing it to move beyond conceptual critique toward a more rigorous and implementable governance framework. The deliberative process helped us identify weaknesses, clarify ambiguities, and refine the proposal into something more balanced, scalable, and institutionally credible. It also ensured that the framework addressed practical realities such as faculty burden, language differences, institutional variation, and evaluative safeguards rather than remaining purely theoretical. Most importantly, the discussions and critiques helped sharpen our central insight that the real crisis is not AI itself, but the continued reliance on assessment systems that infer understanding from submitted outputs rather than requiring it to be demonstrated directly.

Video pitch

https://drive.google.com/drive/folders/17iQKFJL0fRnPTWho09t9okEinYdMZxZd