Why This Toolkit?
AI will force us humans to double down on those talents and skills that only humans possess. — David Brooks, The New Your Times (Brooks, 2023)
AI tools can explain a hard concept, quiz you before an exam, or push back on an argument you are testing. They can also do your thinking for you, and when that happens, the learning stops.
Using AI to learn is different from using it to get a task done. A shopping list or a routine email rewards speed. Learning rewards time, practice, feedback, and some struggle, the same things you get from working with classmates and instructors. The University of Toronto’s student guide warns, “If using AI made the learning process feel too easy, that might indicate that durable learning didn’t occur” (University of Toronto).
The research points the same way. In a field experiment with nearly a thousand high school math students, those who practiced with an unrestricted version of GPT-4 scored 48% higher on practice problems. When the tool was taken away for the exam, they scored 17% lower than students who never had it. A tutor version that gave hints instead of answers largely avoided that drop (Bastani et al., 2025). In an MIT preprint that has not yet been peer-reviewed, students who wrote essays with ChatGPT struggled to quote from essays they had just submitted (Kosmyna et al., 2025; see Stankovic et al., 2025, for a critique of its methods).
UVA’s guidance says AI “may support learning, but it must not substitute for the intellectual work an assignment is designed to develop or assess” (Guidance for Teaching in a World with GenAI, FAQs). This toolkit helps you use AI on the right side of that line.
The toolkit is for students in every UVA school, undergraduate, graduate, and professional. Its examples come from across the University. Your instructor’s rules for each course come first.
What AI Literacy Means at UVA
UVA Library’s AI Literacy and Action Lab grounds its work in a framework developed by Leo S. Lo, UVA’s dean of libraries. The framework describes AI literacy in five domains. Each domain has four levels, from basic awareness to strategic insight (Lo, 2025a, 2025b). The six steps in this toolkit build all five.
Domain |
What it asks of you |
Where this toolkit covers it |
|---|---|---|
|
Ethical awareness |
Recognize issues of bias, fairness, privacy, and accountability, including your obligations under the Honor Code. |
Steps 1, 2, and 6 |
|
Technical knowledge |
Understand, at a basic level, how AI tools produce answers and where they break down. |
Step 2, “How these tools work” |
|
Critical thinking |
Question AI output, check it against evidence, and trace claims to their sources. |
Step 2, “Think critically” |
|
Societal impact |
Consider how AI affects work, equity, culture, and the environment. |
Step 2, “Consider the wider impact,” and Step 4 |
|
Practical skills |
Use AI tools well and recognize when human judgment should take over. |
Steps 3, 4, and 5 |
The sources cited on this page are listed in the Toolkit References.