Shannon Library·University of Virginia
Introduction to Storytelling Across Cultures
Using AI Without Outsourcing Your Thinking
Research + AI · A Guide for Humanities Students
When does AI support your thinking — and when does it replace it?
Fluency is not the same as reliability.
AI may be able to generate language quickly, but you are still responsible for interpretation, verification, and argument. That part stays with you.
Part 2 · What is the intellectual work?
- AskTurn a topic into a question. A topic is not yet a research question.
- FindLocate peer-reviewed scholarship: sources participating in a scholarly conversation.
- ReadRead for interpretation: patterns, assumptions, tensions, contexts.
- CheckVerify. Slow down and test what you have found.
- ArgueMake your own claim using the evidence you have found and read.
Assistance helps your process.
A tool helps you with a process that you still control and evaluate.
Substitution replaces your thinking.
You hand over the intellectual work the assignment is asking you to practice.
Which part of this research process do you think students are most tempted to outsource — and why?
Part 3 · Prompt Makeover Lab
“Write my paper about trickster tales across cultures.”
- question formation
- source discovery
- scholarly reading
- interpretation
- evaluation
- the student’s own argument
“I’m beginning research on trickster tales across cultures. Help me brainstorm 4 possible research questions exploring how tricksters challenge authority. For each, suggest what kinds of scholarly evidence I would need to look for. Do not write an argument.”
- Which question is actually interesting?
- Which question is narrow enough for the assignment?
- What cultural traditions or story versions will they focus on?
- What keywords should they search?
- Which peer-reviewed sources are real and relevant?
- What do those sources actually argue?
- What claim can the student make based on the evidence?
- What learning does Prompt A skip?
- What intellectual work remains yours in Prompt B?
Part 3 · Prompt Makeover Lab
The difference is not just wording. The difference is where the intellectual work lives.
When you use AI, design your prompt so that it helps you move to the next stage of research, rather than jumping to the final product.
- “Help me brainstorm possible research questions.”
- “Suggest keywords I could try in a library database.”
- “What kinds of scholarly evidence would I need?”
- “What are different angles I could explore?”
- “What questions should I ask while reading this source?”
- “Write my paper.”
- “Give me my thesis.”
- “Summarize these articles so I do not have to read them.”
- “Tell me what this source proves.”
- “Create citations for my bibliography.”
Am I using AI to help me think, or am I asking AI to think for me?
Part 3 · Prompt Makeover Lab If time remains
The same move, for Cinderella retellings
“Explain what Cinderella means in different cultures.”
That sounds reasonable, but it invites a broad, flattened answer.
“I’m comparing two Cinderella retellings from different cultural contexts. Help me identify possible analytical lenses, such as gender, family obligation, class, migration, or adaptation. For each lens, suggest what kinds of scholarly sources I would need before making an argument.”
Again, the AI is not producing the interpretation. It is helping the student prepare to do the interpretation.
Part 4 · Interrogating an AI claim
“Trickster tales universally function as resistance narratives in colonized cultures.”
Everywhere according to whom? Based on what evidence?
Which cultures? Which colonial histories? Which time periods?
No specific tale. No scholar. No article. No primary source. No historical context.
- What words or assumptions should make us pause?
- What evidence would we need?
- What context might be missing?
AI can flatten nuance.A possible starting hypothesis. Not something to copy into a paper as established fact.
Part 5 · The verification routine
Stop. Find. Read. Verify.
- StepSTOP
Treat fluent claims as unverified hypotheses.
- StepFIND
Look for real sources in Virgo, JSTOR, MLA International Bibliography, and other library tools.
- StepREAD
Read the primary story and secondary scholarship directly.
- StepVERIFY
Does real evidence support the claim, or is it too broad?
AI gives you a citation? Find the source.
AI gives you a claim? Verify the claim.
AI gives you an interpretation? Think through it yourself.
Part 5 · Verification and the library
“Trickster tales universally function as resistance narratives in colonized cultures.”
“What do scholars say Anansi stories mean in West Africa and the Caribbean?”
Or: “Is Anansi always a rebel, or only in some stories?”
That is what verification does. It does not just catch errors. It improves your thinking.
- If AI gives you a citation and you cannot find it, bring it to us.
- If AI gives you a broad claim and you need real scholarship, ask us.
- If you are not sure whether a source is peer-reviewed, ask us.
- If you are not sure whether a folktale itself is a primary source or whether an article about the folktale is a secondary source, ask us.
The library is where you learn how to trace claims back to evidence.
A claim is not reliable because it sounds confident.
A citation is not real because it looks real.
Part 5 · Verification If time remains
This looks real. That doesn’t mean it exists.
Vanchev, Mira. “Trickster Resistance in Global Folklore.” Journal of Folklore Studies, vol. 12, no. 3, 2003, pp. 45–67.
- Does this article title appear in Virgo or a library database?
- Does the journal exist?
- Does the author exist and publish in this field?
- Is there a DOI or stable record?
- Can I access the full text?
If I cannot verify it, I cannot cite it.
This citation was invented for the exercise. AI can create citations that look academically formatted but are completely fabricated.
Part 6 · AI-use spectrum
Rank these from least intellectual work outsourced to most intellectual work outsourced:
- Scenario 1
“Give me 10 keywords I could use to search library databases for Cinderella adaptation scholarship.”
AssistanceYou expand search vocabulary, but you still use library tools, find real sources, read them, and decide what they mean.
- Scenario 2
“Give me 5 peer-reviewed journal articles analyzing Chinese Ye Xian Cinderella variants.”
High riskAI can invent citations. Every single source must be verified in library systems before you trust it.
- Scenario 3
“Read these two articles and tell me what thesis statement I should write for my class.”
SubstitutionThe clearest example: developing a thesis from evidence is the central intellectual work of the assignment.
- Scenario 4
“I read this article on West African trickster tales. Help me brainstorm 3 counterarguments to the author’s main claim.”
AssistanceYou have already done the reading; AI acts as a sounding board. You still evaluate whether the counterarguments make sense.
What work is being handed over?
If AI helps you generate vocabulary or possibilities, it may be assisting your process. If AI reads, interprets, decides, or argues for you, it is replacing the work you are here to learn.
Part 7 · Responsible-use checklist
Before using AI, ask:
AI can help you research. It cannot be responsible for your research.
Part 8 · Exit ticket
3-Line Exit Ticket
Research in storytelling is about making meaning across diverse human cultures. AI can help you get unstuck, but it cannot care about cultural context, it cannot read for nuance, and it cannot make an argument on your behalf.
That work is yours — and the librarians here at Shannon Library are here to help you do it.
Thank you.Questions?
Using AI Without Outsourcing Your Thinking · Shannon Library