Acknowledging AI

When they imagine students using AI, faculty might picture a student plugging an assignment prompt wholesale into ChatGPT and submitting the output with minimal or no editing. But students—who are far more prolific users of AI than faculty—report that this approach to AI use is relatively uncommon. AI tools are more often leveraged as collaborators that enhance students’ writing, rather than replacing their efforts wholesale. As reflected in UMN SERU data, generative AI tools play a variety of collaborative roles in students’ writing: suggesting and summarizing sources, assisting in brainstorming and outlining, explaining difficult course material, providing general feedback or line edits, generating images or sound, writing and debugging code, and more.

In TWW workshops and consultations with faculty and instructors, we hear repeatedly that they want to help students use AI tools in critical ways that enhance their work without detracting from their thinking, learning, and agency. To start these meaningful conversations, we need ways for students to capture, share, and reflect on the full range of ways they use AI tools. This blog proposes acknowledgment assignments as one useful option.

How Do We Acknowledge AI Use?

When considering how students should acknowledge AI use, many faculty default to the most common and visible form of academic attribution: citations. This approach is incorporated in the UMN Faculty Senate syllabus statements on AI use and aligns with recommendations from major style guides (e.g., APA, Chicago, and MLA).

But while some uses of AI are fairly easy to acknowledge with traditional citations, many of the collaborative uses of AI employed by students leave the boundaries between the student’s and the AI tool’s language and ideas somewhat nebulous. These uses of AI more closely resemble the ways we might use reference materials, word processors, citation generators, databases, or even conversations with colleagues than the ways we use sources.

Fortunately, academics have multiple ways of documenting the contributions of others. For example, authors regularly recognize those who supported their writing in acknowledgment statements. Acknowledgments share and express gratitude for the insights and contributions of others. In so doing, they often offer glimpses of a writer’s process and reveal the social, collaborative nature of even solo-authored writing.

As they respond to the ongoing evolution of AI tools, numerous academic journals now ask authors to disclose their use of AI tools in their acknowledgments sections. Similarly, some UMN faculty are exploring acknowledgments assignments as sites for students to share and reflect on their writing process, including their use of AI tools. Through acknowledgments, students can share how AI tools have shaped their writing process and product in ways that can’t be neatly captured by citations. Providing such an acknowledgment not only improves the accuracy of disclosure, but also sets the stage for thoughtful reflection on AI use.

Example: Dr. Molly Ball’s Acknowledgments Assignment

Dr. Molly Ball in American Studies has developed an acknowledgments assignment for her writing-intensive American Popular Culture and Politics course. For each paper in the class, students write a 300-word statement identifying “the people, tools, sources, and more that helped” them develop their writing, as well as describing and reflecting on the broader writing process that leveraged those resources. This assignment is a place for students to acknowledge AI use, but that’s not the exclusive focus. Instead, students are asked to share and reflect on all the sources of support they received, whether or not they chose to use AI.

If they’ve chosen to use AI, students are asked to disclose details like:

• What AI tool(s) they used (e.g., CoPilot, ChatGPT, Claude, Grammarly)
• What prompts they gave to the AI tool(s)
• Ways AI tools influenced their process (e.g., for brainstorming or summarizing sources) and final written product (e.g., grammar correction or supplying a counterargument for the student to rebuff)

As Dr. Ball explains to her class, this assignment is at its core a metacognitive activity, one where students can identify effective and ineffective practices in an effort to refine their writing process and AI literacy throughout the semester. Because the assignment is low-stakes—students are guaranteed at least an 85% if they answer the basic questions in the assignment, and there is no ‘wrong’ process for which they could lose points—it provides space for honest exploration.

Dr. Ball’s assignment is similar to a process memo or revision memo, a genre some instructors may have used previously in their  courses. We shared strategies for developing reflective assignments in a previous TWW blog.

Maximizing the Effectiveness of Acknowledgment Assignments

If you are interested in designing your own acknowledgment assignment, consider these strategies to maximize its effectiveness:

• Don’t make disclosing AI use additional work. Requiring only students who used AI to write a lengthy statement penalizes students for using available tools—suggesting that avoiding AI is the ideal—and encourages them to hide their AI use to avoid extra work. A well-designed acknowledgment assignment asks all students to acknowledge and reflect on their choices in the writing process, whether or not they chose to use AI.
• Remind students what tools count as AI. For instance, many students are not aware that Copilot assistants automatically suggested by their browser or some versions of Grammarly are AI tools.
• Clarify which practices should be acknowledged. After all, even conducting a Google search now involves using an AI tool—does every Google search related to the project need to be cataloged?
• Connect to learning outcomes. Connect your AI policy for assignments to learning outcomes, considering how AI tool use may affect student learning, both positively and negatively. Similarly, ask students to connect their reflection to assignment learning outcomes: how did their choices around process and AI tool use enhance or detract from their learning?
• Make it clear that disclosed AI use will not be penalized—and keep your word. Students are often quite anxious that disclosing AI use may lead to their work being devalued or even lead to academic sanctions for plagiarism, and there is evidence to support their concern. Consider steps to mitigate bias, such as reviewing students’ acknowledgment statements anonymously or after grading the accompanying papers.

Thoughtful acknowledgment assignments like Dr. Ball’s set the stage for students and faculty to reflect together about how AI tools are changing writing in their courses and disciplines. If you’d like support in developing your own acknowledgments assignment, request a TWW consultation. We invite you to comment below to share how you’re thinking about acknowledging AI use.