AI Use Guidelines for Courses and Research
Department of Statistics, University of Michigan
April, 2026
The AI revolution is rapidly changing both the practice of statistics and data science, and how these subjects are taught. To support students and instructors, the Department of Statistics supports the following principles for AI use in our courses.
1. All courses should have learning goals that aim to build foundational knowledge, which students should master and be able to apply autonomously. This supports their intellectual development, and enables them to critically evaluate AI-generated material.
2. Many courses will also have learning goals that are appropriate for AI assistance. These will generally be related to practical skills that are secondary to the main focus of the course.
3. Students should be informed of the rationale for identifying learning goals as being foundational or suitable for AI assistance. This will vary by course. For example, writing computer code without AI assistance may be a foundational goal in one course, but in another course it may be appropriate for students to use extensive AI support in their coding.
4. If a course has both foundational and AI-assisted learning goals, homework should help students achieve both. Students should be informed which homework tasks are intended to build foundational skills and will later be assessed in an environment where AI is not available.
5. Given the difficulty of detecting inappropriate use of AI, courses should be designed so that students cannot gain an unfair advantage in their course grade by using AI for tasks where it is prohibited or discouraged. Tests taken without AI assistance should carry enough weight so that course grades primarily reflect students’ mastery of foundational skills.
6. If a course includes learning goals for which AI use is appropriate, students should be taught and assessed on the effective and responsible use of AI.
In addition, the Department of Statistics endorses the following guidelines for use of AI in theses and other research works used to meet degree requirements.
1. Students must be able to independently explain and defend (without using AI) all aspects of their research products in depth and with rigor.
2. Any errors, limitations, or shortcomings of the research work are solely the responsibility of the student authoring the work.
3. Any research work employing AI should include a clear statement describing how it was used, ideally including steps taken to mitigate the risk of misleading or erroneous AI-generated content.
4. Students and their mentors are expected to discuss the role of AI in the research, as well as how its use is documented in the thesis, in the context of current standards for responsible scientific research.
5. Theses should reflect the unique voice and communication style of the student authoring the thesis.
