GenAI Assessment Scale for BA in ELT Programs

GenAI Prompt

Role: Act as an expert in curriculum instructional design and GenAI policy for higher education.

Context: I am an English language teacher trainer operating within a BA program. We have a dual mandate: developing students’ General and Academic English proficiency while simultaneously equipping them to become future educators.

Source Material: Below, I have provided our original 5-level GenAI Assessment Scale (the core AIAS).

Task: I need you to adapt this general scale into three distinct, specialized versions for different strands of our BA-ELT curriculum:

1. English language skill development subjects

2. English for Academic Purposes (EAP) subjects

3. Teaching practicum subjects

Directives & Formatting:

• Introduction: Start with a brief paragraph explaining why a hybrid framework is optimal for our dual mandate, mentioning the need for students to master the pedagogical orchestration of AI (GenAI Deep Research).

• Adaptations: For each of the three subject areas, include a “What changes” summary explaining the shift in assessment targets and AI constraints. Then, adapt Levels 1 through 5 for that specific context.

• Pedagogical integration: When drafting the Teaching Practicum section, integrate concepts like SLA and UDL into the AI usage examples.

• Implementation Warning: After the three scales, write a paragraph warning faculty against “retrofitting” old rubrics, explaining how rubrics must fundamentally shift to assess AI interaction (e.g., prompting, digital literacy) at higher levels.

• Archive: Finally, append the original 5-level scale I provided at the very bottom under the exact heading: “General (original) GenAI Scale (Archived)”.

[Insert Original 5-Level GenAI Scale Here]

For an English language teacher trainer operating within a BA program—where the dual mandate involves developing students’ general and academic English proficiency while simultaneously equipping them to become future educators—the optimal assessment framework is a hybrid integration of the core AIAS and the EAP-AIAS. This specialized BA-ELT GenAI Assessment Scale ensures learners develop internal linguistic schemas while mastering the pedagogical orchestration of AI (GenAI Deep Research).

GenAI Assessment Scale for English language skill development subjects

Adaptation for skill-development courses (General English, integrated skills, pronunciation, speaking/listening/reading/writing skills)

What changes: keep the assessment target on language growth and communicative performance; allow AI mostly as input/support, and require evidence of learning (not just a polished product).

• Level 1: No AI. Best for timed writing, live speaking tasks, in-class reading/listening checks, and baseline diagnostics.

• Level 2: AI for comprehensible input + planning (graded texts, glossaries, models, practice items). Student output remains fully human.

• Level 3: AI as a coach (feedback, error noticing, reformulation options). Require a short “change log” explaining which suggestions were used and why.

• Level 4: AI supports complex performance tasks (e.g., extended presentations, debate prep, integrated reading-to-write). Grade the student’s verification, selection, and revision of AI help + accuracy and appropriateness for the target level.

• Level 5: AI integrated into personalized learning systems (e.g., students build a weekly micro-syllabus, spaced-repetition deck, or speaking practice plan). Grade the quality of the learning design + reflection on progress and transfer to real communication.

GenAI Assessment Scale for English for Academic Purposes (EAP) subjects

Adaptation for EAP courses (academic reading, writing, research literacy, citation practices)

What changes: increase the emphasis on research integrity, evidence handling, and genre control; require documentation of sources and AI use at every level above 1.

• Level 1: No AI. Best for in-class academic paragraphs, source-use quizzes, summary/paraphrase under supervision, and timed argument writing.

• Level 2: AI for finding, clarifying, and planning (concept checking, outline options, question refinement). Student drafts are human; require an “AI planning appendix” (prompts + what was kept/discarded).

• Level 3: AI as editor + critical friend (clarity, cohesion, hedging, register). Require a revision trail + academic integrity statement (no fabricated references; verify all claims).

• Level 4: AI supports research tasks (literature mapping, gap statements, methodological options). Grade the student’s fact-checking, source verification, citation accuracy, and substantive rewriting (not surface-level edits).

• Level 5: AI integrated into advanced scholarly workflow (synthesis across many papers, building a theoretical frame, designing an original study/argument). Grade the student’s synthesis quality, epistemic caution, and strategic prompting + transparent documentation.

GenAI Assessment Scale for teaching practicum subjects

Adaptation for teaching practicum courses (lesson planning, materials design, microteaching, reflection, assessment design)

What changes: keep the assessment target on pedagogical reasoning and classroom decision-making; require alignment (objectives–tasks–assessment) and ethical use (privacy, bias, accessibility).

• Level 1: No AI. Best for live teaching performance, observation notes, and spontaneous classroom decisions.

• Level 2: AI for idea generation + planning support (activity options, instructions, differentiation ideas). Require the student to justify choices using course principles (e.g., SLA/UDL).

• Level 3: AI as a design partner (worksheet drafts, rubrics, feedback phrasing). Require annotated revisions showing how the student adapted materials for learners and context.

• Level 4: AI supports complex professional tasks (unit plans, assessment systems, intervention plans). Grade the student’s evaluation of AI suggestions, feasibility in the real classroom, and evidence-based rationale.

• Level 5: AI integrated into advanced teacher expertise (data-informed reteaching cycles, curriculum innovation, inclusive design at scale). Grade the student’s strategic orchestration of AI across planning–teaching–assessment–reflection, with clear safeguards and documentation.

When implementing this scale across the BA curriculum, it is imperative to avoid “retrofitting” old assessments. The rubrics must be fundamentally updated to align with the chosen level. For instance, while a Level 1 rubric focuses entirely on the student’s independent demonstration of knowledge and skills, Level 2 and Level 3 rubrics must include explicit criteria evaluating how students engaged with and modified AI outputs. Furthermore, Level 4 and 5 rubrics must assess overall AI literacy, including the ability to prompt effectively and integrate AI into a broader academic workflow.

General (original) GenAI Scale (Archived)

Level 1: Independent Language Production (No AI Use)

• Focus: Developing and assessing core language skills and academic competencies entirely independently and without AI assistance.

• Application: Utilized for establishing authentic baselines in reading, writing, and speaking. Future teachers must experience the cognitive friction of language acquisition without technological intervention. Tasks are completed without the use of technology and include real-time oral debates, spontaneous in-class presentations, etc.

Level 2: AI-Assisted Language Input and Planning

• Focus: Using GenAI as a highly sophisticated tutor for pre-task activities, schema building, and finding and understanding literature, while final text production remains human.

• Application: In General English and skill-based subjects, students might use AI to lower the lexical level of authentic texts or generate bilingual glossaries for example. In EAP, learners review AI-generated models of academic genre conventions, self assessing reading and listening comprehension, etc. before drafting their own independent essays. Assessments require students to explicitly show how they refined these initial AI-generated ideas, often via interim checkpoints or planning logs.

Level 3: AI for Limited Language Practice and Collaboration

• Focus: Humans and AI co-create text, with the AI acting as a “critical friend” for drafting, obtaining feedback, and refining discipline-specific collocations.

• Application: Highly effective for intermediate writing modules. BA ELT students draft an initial text and feed it into an AI to check for academic clichés or grammatical errors. Crucially, assessment at this level mandates a “revision trail”: students must submit their original draft alongside the AI-assisted version, featuring tracked changes and commentary explaining their pedagogical rationale for accepting or rejecting the AI’s suggestions.

Level 4: AI-Assisted Task Completion with Critical Evaluation

• Focus: Students direct the AI to assist in complex academic tasks but are rigorously evaluated on their ability to critically evaluate, fact-check, and substantially revise the outputs.

• Application: Deployed in advanced EAP and research methodology classes. Trainee teachers might use AI for initial literature gap-finding or generating draft sections of a research proposal. The assessment shifts from evaluating linguistic mechanics to grading the student’s digital literacy, prompting precision, and ability to prevent AI hallucinations.

Level 5: Selective AI Integration for Advanced Pedagogical Skills

• Focus: Extensive AI use aimed at enhancing student work and preparing them for real-world, discipline-specific scenarios in their future teaching careers.

• Application: Ideal for capstone ELT research projects or advanced curriculum design modules. Future educators co-design tasks, prompt the AI to synthesize themes across multiple dense pedagogical papers, identify gaps in the AI’s synthesis, and develop their own overarching arguments.