TIPS for Teaching

At the recent USC Upstate AI Summit in September, participants asked what we know about our students’ view of Generative Artificial Intelligence (GenAI). An exploratory survey of USC Upstate students, faculty, and staff was conducted by members of the AI Training Subcommittee of the AI Strategic Council in Fall 2025. In it, 51% of students identified themselves as AI aware, with a basic understanding of tools they were just starting to test out.

Although the sample is too small to represent the entire student population, the responses offer valuable insights into the ways students use, view, and want to learn about GenAI. Faculty can build on what we know about students’ existing uses of GenAI to develop more sophisticated, critical, and ethical AI literacy skills.

How Students Are Using GenAI

The most common ways students reported using GenAI included brainstorming (64%), summarizing or explaining complex concepts (61%), and creating study tools like flashcards or practice tests (57%). Only 11% selected writing papers, though 43% chose outlining papers and 36% indicated they used GenAI to gather research.

Interestingly, 32% noted they had used GenAI for creating resumes, but only 7% used it to complete job applications and only 21% reported using GenAI to assist with interview preparation.

These preliminary results suggest that students who use GenAI are largely employing it to support their learning rather than to replace independent thinking. However, their responses raise questions about whether they have developed more advanced AI literacy skills of critical evaluation, ethical decision-making, and transferring AI skills to professional contexts.

Although many students reported relying on GenAI for foundational materials, like summaries and explanations, outlines of ideas, and research, their greatest concerns centered on the potential consequences of AI use as a basis for learning and thinking:

• 88% were moderately or extremely concerned about accuracy and reliability.
• 83% were moderately or extremely concerned about losing academic skills, including critical thinking and writing.
• 83% were concerned about data privacy.
• 81% were concerned about academic misconduct.

The concerns students expressed about the potential costs of AI use align with key AI literacy principles about how AI systems work, how to recognize bias and hallucinations, how to critically evaluate and verify AI outputs for accuracy and bias, and how to maintain integrity by citing AI assistance.

These responses suggest opportunities to strengthen students’ AI literacy skills, both to make their current GenAI use as effective as possible and to engage in more high-impact, human-centered, ethical practices that can contribute to their academic and professional development.

How Students Want to Learn About GenAI

The core question for faculty becomes how to help students develop critical AI literacy skills. Among students who responded to questions about their experience with GenAI:

• 78% described themselves as just learning about GenAI or having only a basic understanding.
• 58% reported no weekly use of AI tools like ChatGPT, Copilot, Gemini, Claude, or Grammarly.
• 73% had not participated in any GenAI training or learning opportunities.

When the survey asked students what forms of GenAI training they would like to have, 65% said none. These results remind us that students have different levels of interest in AI, and that developing AI literacy should include opportunities to critically examine the technology rather than simply encouraging its use.

Among the training options offered, integrating AI into existing required courses and attending university-sponsored workshops were the most frequently selected, each chosen by 21% of respondents. Students were less likely to select dedicated AI courses, interactive games, or student organizations focused on AI.

More than any other type of support listed on the survey, students selected examples of appropriate AI use as the key to helping them feel more confident using GenAI.

• For example, a faculty member might demonstrate how ChatGPT.Edu can generate practice questions while also showing students how to verify sample answers against course materials. Or better yet, encourage metacognition by asking students to explain the concepts to a chatbot programmed to ask clarifying questions. Students can then reflect on where their explanations were strong and where they revealed gaps in understanding.
• An instructor may assign a research project that begins in class using GenAI in groups, then asks students to locate and verify the sources using the library or other credible sites. Then, show students how to design follow-up prompts to help them evaluate the sources’ credibility, accuracy, and relevance to the project or to find gaps and blind spots in their research.
• Faculty might build on students’ existing use of GenAI for brainstorming by asking students to refine prompts with greater context, solicit multiple possible responses and compare them, and explain which ideas they accepted or rejected and why.

Next Steps Toward AI Literacy

As we incorporate AI into our academic settings, it is important to continue to invite student perspectives on how they can best learn about this technology that is bringing waves of change to society. Consider starting up a conversation rooted in the needs of your discipline and your course. Then, guide students toward higher levels of AI literacy that will help them make critical evaluations and ethical decisions in your course and beyond.

We can start with a simple question:

Some of the greatest productivity benefits of AI can be found in tasks you are performing at the upper range of your skill set. AI can help you level up.

When you know enough to accomplish a task, but not enough to do it confidently or efficiently, AI can provide you with step-by-step guidance to speed up your problem-solving process.

The key is to ask for step-by-step instructions. ChatGPT often explains its reasoning by default, but you can also ask for explanations that meet you at your particular skill level.

An AI Breakthrough Example: Photo Roster

For three years, we’ve been facing the challenge of accessing a useful Photo Roster in Blackboard or Self-Service Carolina. As instructors, we need a page of images large enough to use for class attendance, while preserving the preferred names and pronunciation notes that these systems provide.

After doing some initial experiments, I knew the coding was simple, but not quite simple enough for my skills. No staff member with the right skill set had the dedicated time to provide the solution.

Enter ChatGPT

With ChatGPT, I was able to explain the problem, offer an example of my initial solution, then work iteratively to tweak the code until I had exactly the photo roster display I wanted. I could run a test and share a screenshot of the output to identify the right line to target in the code.

The end result is a Blackboard Photo Roster Bookmarklet available to anyone who accesses Blackboard Ultra courses through a web browser.

For me, javascript coding was the skill I was hoping to “level up,” but the same iterative, step-by-step process can be used for creating complex branching in surveys, automating data collection across Excel spreadsheet sources, or any other task that can be broken down and tested with images of outputs.

Without ChatGPT, I would have to study several more advanced skills to decipher the complete code for the existing roster page. Then, I could apply that knowledge to identify just the areas I needed to accomplish my goal. With ChatGPT, I was able to use GenAI to determine the targets for my very focused changes to the display settings. I can learn the rest of the details when I am ready to level up again.