Skip to Main Content
Armacost Library
Ask Us

SOCIOLOGY & ANTHROPOLOGY: Using Generative AI Well

Generative AI tools like ChatGPT can be powerful partners in your work, but they are not infallible. These guidelines will help you use AI effectively, responsibly, and in ways that strengthen—not replace—your own thinking.

Five Principles

1. Treat it like a person, not a calculator or robot

Of course AI is a machine.

But we have to be careful not to think of it as something like a calculator. Calculators don’t make mistakes. Generative AI absolutely does. Computers are programmed very specifically- they are reliable, dependable, can take input and give output. AI does not- it can be “creative.”

So in that sense, it makes sense to treat it like a person. People aren't always reliable, they don’t have all the facts, etc. But you still ask them for information, advice, feedback, and trust them to a certain extent. 

So while this feels strange to say, you ought to treat it more like a person than a machine.

2. Learning and knowing
AI does not “know” things the way people do. It is extremely good at producing plausible text, not at being an expert. Do not use it as your primary tool for researching topics or for fact-finding. If you need scholarly sources, use library databases and catalogs, then read and evaluate the articles yourself. AI can help you think through material you already understand, but it can also sound confident while being wrong. Treat any “facts” it gives you as unverified until you confirm them with credible sources.

Side note: this is the reason that chatbots are censored, aka "having guardrails." If a prompt is blocked, reframe your request toward your legitimate learning or communication goal.

3. No job is too weird: some good use cases
AI is best when you are asking it about topics you are already familiar with, because you can do fact-checking in real time. As long as you understand that AI is not always accurate, here are a few possible tasks you could experiment with.

  • Getting unstuck at the start of a project, or when you feel overwhelmed
  • Brainstorming angles, outlines, titles, or next steps
  • Talking through ideas to clarify your thinking
  • Drafting tricky emails or announcements
  • Asking for plain-language explanations of technical passages
  • Spotting what you might be missing in a plan or argument
  • Generating high-level overviews, report skeletons, and to-do lists (even in very fine detail)
  • Drafting short pieces such as invitations, announcements, or form letters
  • Language support if English isn’t your first language (still review carefully)
  • Rough translations to get the gist (use with caution if you don’t know the language)

ChatGPT is especially famous for just agreeing with you, so you'll want to also ask it to be critical, if that's what you want.

And in all cases, you remain responsible for the accuracy, tone, and appropriateness of any final product you use it for.

4. It doesn’t judge you (use that to your advantage)
Although you’re treating it like a person, AI is a robot, which is good! Because AI doesn’t judge you (unless maybe you ask it to?)

AI is very good at taking your rambling thoughts and bad notes and organizing them and/or doing something with them. Before you dump your notes in, you can also give it details about who you are, what you’re doing, who you are delivering a completed product to.

5. Only one of you has a brain
It's you! You have the brain!

Read AI outputs carefully. Keep what’s useful, fix what’s off, and discard the rest. Evaluate claims, check facts, compare against your assignment goals, and ensure the final product reflects your own understanding and voice.

A few final thoughts

Academic integrity and responsible use

  • Follow your instructor’s policies on AI-- check the syllabus for each class.
  • When permitted, be transparent about how you used AI (e.g., “outline brainstorming,” “first-draft phrasing”), especially for graded work.
  • Be especially cautious when asking for citations or sources. It's best to find sources from the library or Semantic Scholar (or similar tools).
  • Protect your privacy. Avoid putting confidential, personal, or proprietary information into prompts.

Bottom line
Use AI to think with—not to think for—you. It can jump-start ideas and tidy drafts, but you are the one who learns, decides, and is accountable for the final work.

Author

Emily Croft
Systems and Technology Librarian
emily_croft@redlands.edu

Adapted from a talk delivered to URSA, April 2, 2025
last updated: August 14, 2025