GAI Use Scale a tool for supporting communication between instructors and students

A Tool to Enhance Communication on GAI and Learning

Instructors play a vital role in guiding students to learn about generative AI (GAI) and discern how to use it effectively for specific purposes. A 2024 Student Voice survey from Inside Higher Ed, for example, asked a nationally representative sample of over 5,000 students if they "have a clear sense of when/how/whether to use generative artificial intelligence to help with [their] coursework," to which the top two responses—a tie—were "Yes, because some or all of my professors have addressed this issue in class" and "Not sure." In other words, whether an instructor addresses GAI substantively can make the difference between a student feeling prepared or unprepared to engage with it in their learning and beyond.

There is no single threshold policy for GAI in coursework, and disciplinary nuance adds further variety to how GAI—itself a broad category that includes different capabilities and functions—aligns with the aims of specific courses and instructors. Moreover, some activities and assignments may call for deeper engagement with GAI while others in the same course may not.

To provide a vocabulary for students and instructors to communicate about expectations for GAI, CELT has developed the Student AI Use Scale from established precedents in GAI literacy frameworks, GAI assessment scales, and emerging evidence on student learning with GAI. Instructors may use the levels of the scale to label individual assignments (e.g., in prompts or instructions) or their course more broadly (e.g., in the syllabus) to clarify expectations for student learning.

The scale also may be adapted to the specific nature of a course or discipline, as well as to an instructor's unique goals and expertise. The four levels on the scale represent a progressive increase in possibilities for GAI use to empower students as discerning users and active learners as they engage with this new technology.

It's important to note that the GAI Use Scale is part of a larger set of strategies to clarify expectations and opportunities for GAI use in individual courses and assignments. Instructors may contextualize the scale in syllabi and other documentation with further explanations of their GAI policy as detailed in the UK ADVANCE guidelines.

Student AI Use Scale

GAI may not be used for any reason. All steps taken for the assignment are completed by the student with no assistance from GAI.

When completing assignments and tasks identified as this level, students are responsible for generating all core ideas, content, prose, and the overall structure of the work. If allowed by the instructor, a student may use generative AI for very specific, limited tasks like proofreading, but the work submitted represents the student’s original thoughts, ideas, and labor. Any use of AI on the student’s part, if allowed, likely happens at the end of the assignment process after the student has completed the initial efforts.

GAI’s Role in Level 1 Assignments

AI plays the role of reviewer or limited editor at this level. While it may act as a helper for basic editing of the student’s pre-existing work or finding generalized definitions, an assignment labeled this level is done almost exclusively by the student.

Examples of allowable student use of GAI at level 1. Use GAI for . . .

Suggesting minor edits for improving clarity.
Proofreading for common grammatical errors.
Reviewing citations for adherence to a particular style.
Critiquing data analysis methods to inspire student-centered revision to approaches.
Generating a practice quiz for study help.

When completing assignments and tasks labeled level 2, students are responsible for the driving ideas, conceptualization, and development of the work being completed. However, students can actively use generative AI to brainstorm, elaborate, or expand on their original ideas. They can also use GAI to help invite different perspectives, create organizing structures, or problem-solve. Any use of GAI should be critically evaluated, refined, and integrated thoughtfully into the student’s own original thoughts and arguments. Student use of GAI at this level is meant to provide initial sparks in thinking and progress in the conceptualization and development stages.

GAI’s Role in Level 2 Assignments

GAI serves as a thought partner for ideation, organization, or brainstorming. While the student is still the ultimate driver behind procedures/processes, writer of prose, and creator of content, GAI is sprinkled into the task in strategic places to improve the final product or solve challenges.

Examples of allowable student use of GAI at level 2. Use GAI for . . .

Brainstorming paper topics or narrative approaches.
Designing recommended outlines for delivering information.
Generating ideas or approaches.
Designing examples for inspiration.
Comparing different research approaches.
Troubleshooting or identifying potential errors in student-authored code.

When completing assignments labeled level 3, students are responsible for collaborating with generative AI to perform various tasks. Students can call on GAI tools to help them complete tasks like drafting content, examining GAI outputs, critically refining GAI-generated content, designing content outlines, or performing structured ideation for content co-development. Students using GAI at this level typically do so in service to larger, more complex projects and will participate in back-and-forth engagement with GAI tools. Student use of GAI at this level typically happens throughout the assignment processes.

GAI’s Role in Level 3 Assignments

AI serves as a co-creator for drafting and generating substantial content segments with student guidance.

Examples of allowable student use of GAI at level 3. Use GAI for . . .

Generating an initial draft of a literature review or background section based on provided sources and an outline.
Drafting specific paragraphs or sections of a report, essay, or creative piece, which the student then substantially revises for voice, argument, and accuracy.
Co-designing a research methodology section by prompting GAI for different approaches and integrating its suggestions into their own refined plan.
Producing code snippets or functions that solve specific parts of a programming problem, which the student then critically evaluates, integrates, and debugs.
Developing different argumentative structures for a persuasive essay, with the student choosing, adapting, and filling in the core arguments.
Creating a preliminary marketing brief based on student-provided core information, with GAI assisting with generating some structure and content.

When completing assignments labeled as level 4, students are responsible for leveraging generative AI frequently to solve complex problems, synthesize information, and develop innovative solutions. At this level, students can use GAI to grow more familiar with the technology broadly as well as specific tools as part of creative, intellectual, and problem-solving processes. As with any use of GAI, students must still critically analyze, critique, and edit any GAI outputs they plan to use in their final submissions. However, assignments labeled this level invite students to blend GAI into most, if not all, stages of complex intellectual tasks. This stage of GAI use is recommended only for students who have thorough experience with GAI and a strong base of GAI literacy.

GAI’s Role in Level 4 Assignments

AI serves as an engine for generating solutions, complex content, novel insights, and other contributions to student work as it is guided by the student’s advanced, knowledgeable direction.

Examples of allowable student use of GAI at level 4. Use GAI for . . .

Directing GAI to simulate multiple complex scenarios (e.g., in an economic model, a scientific experiment, or an engineering design) and then critically interpreting and synthesizing the results to draw conclusions.
Using GAI tools to generate and evaluate a wide array of potential solutions to an ill-defined problem (e.g., sustainable energy solutions for a city, or a new ethical framework for an emerging technology), selecting the most promising, and then refining them.
Employing GAI to synthesize information from diverse and extensive datasets or knowledge bases, and then building an original argument or creating novel insights based on that synthesis.
Managing an GAI-driven creative project where the student provides high-level artistic direction and curates GAI-generated content (e.g., images, text, music) into a cohesive, original collection.
Developing and testing multiple GAI prompts to generate comprehensive research outlines, then critically combining and augmenting the best elements into a capstone project proposal.
Designing and implementing a large-scale GAI-assisted programming project, where GAI generates significant portions of the codebase, and the student's role is primarily architecture, integration, debugging, and quality control.

GAI Scale Visual Resources

An image showing the "Student AI Use Scale" developed by the Center for the Enhancement of Learning & Teaching at the University of Kentucky. The scale has five levels, from 0 to 4, each with a description of the student's role and the AI's role. Level 0 is "Sole Author," where the student works independently and the AI's role is "None." Level 1 is "Primary Creator," with the AI as a "Reviewer or limited editor." Level 2 is "Conceptual Architect," where the AI is a "Thought partner." Level 3 is "Critical Co

GAI Use Scale

A chart from the UK Center for the Enhancement of Learning & Teaching titled "WHEN CAN I USE GEN AI?". It outlines five levels of student AI use, from 0 to 4. Level 0, "Sole Author," involves no AI use. Level 1, "Primary Creator," allows AI for proofreading, feedback, and outlining at the end of the process. Level 2, "Conceptual Architect," permits AI for brainstorming and troubleshooting in the initial stages. Level 3, "Critical Collaborator," adds synthesis, reading, and drafting segments, allowing for AI

When Can I Use GAI? Student Resource

How To Use the GAI Use Scale

To implement the scale, instructors should:

identify the appropriate level for their course (or for individual assignments),
provide specific guidance that matches that level for each assignment, and
communicate with students early and often about expectations for GAI use.

To make the best use of the scale for their learning, students should:

examine the levels and their course's assignment prompts/instructions closely,
ask questions early and often,
follow all guidelines in the scale and in specific instructions, and
critically engage GAI tools (e.g., to develop awareness of capabilities and limitations) to become more informed users.

Instructors can use their expertise to identify the key skills, mental moves, and habits of mind that students should acquire in their courses as well as the specific assignments and tasks that leads students to build those skills. Then, instructors may consider where GAI enhances learning and skill development as well as where (and how) it may present a barrier to students' ability to learn and grow.

The student-centered language of the scale (a unique characteristic among scales that typically focus on GAI tools themselves and what actions are permitted) is intended to open up space for dialogue and community building. Further, this student-centered framing acknowledges that an exhaustive account of granular learning tasks and effort for all possible settings quickly becomes overwhelming, ineffective, and unfeasible. Through educational experiences guided by instructors who are experts in their specific fields, students can learn the lasting skill of how to determine effective uses of GAI when the tools are constantly changing.

Other examples of GAI use scales are the AI Assessment Scale (Furze, et al.), the Seven Levels of Possible GenAI Usage (ISU), and the AI Friction Scale (Beckwith).

Important Note for All Levels

Any use of GAI requires rigorous evaluation of both specific outputs and overall influence on process and thinking. All GAI outputs can be improved and most need revision to be effectively integrated into a specific task or situation. Any student using GAI remains accountable for the accuracy, originality, and intellectual integrity of their work. Students should also attend to how GAI use may augment or detract from their own arguments, voice, approaches, and relevant academic or professional standards.