AI for your major

How professionals in your field are using AI right now. About 1 h, intermediate level.

You'll be able to

• Find three real uses of AI in your field and tell them apart from demos and marketing.
• Identify the privacy, liability and professional rules that shape how your field can use AI.
• Try a small version of one workflow and judge what you'd need before you trusted it.

Real uses versus demos

Most of what you'll find about AI in any field is demos: a tool doing something impressive once, on a carefully chosen example. A real use looks different. It usually has:

• A named person or team doing the work, not just a vendor describing it.
• A specific task inside a larger job, like drafting a first pass or sorting documents.
• A check: someone qualified reviews the output before it matters.
• Stated limits: what it's bad at, and what happened when it was wrong.

Good places to look: your field's professional associations and trade publications, talks by practitioners, your instructors and career advisers, and the Hub's video library filtered by your field. Treat vendor marketing as a list of claims to check, not evidence.

Aim for three real uses. If you can only find demos, that's a finding too: it may mean your field is still early, or that the tools don't fit its work yet.

Patterns that repeat across fields

The tools differ by field, but the jobs they do are surprisingly similar. Clinicians use AI to draft visit notes and summaries. Lawyers use it to review contracts and find relevant clauses. Marketers test variations of copy. Engineers and scientists generate code for simulations and analysis. Journalists and social scientists transcribe interviews and do a first pass at sorting open-ended responses.

Underneath, most real uses are one of a few tasks:

• Summarize a long record, transcript or document.
• Draft a first version for a professional to edit.
• Classify or extract information from many documents.
• Write code to analyze or simulate something.

And almost all of them share one structure: AI produces a draft or a sort, and a qualified person checks it and takes responsibility. When you read about a use in your field, look for that person. If nobody checks, ask what happens when the output is wrong.

Your field's rules

Every field has rules that decide what AI use is acceptable, and they matter more than what the tool can do.

• Privacy: health information, student records (covered by FERPA in the US), client files and personal data often have legal protection. Pasting them into a general AI tool can break those rules.
• Research ethics: studies with human participants follow an approved protocol, usually through an IRB, which can limit which tools may touch the data.
• Liability: the professional stays responsible for the work. There have been widely reported cases of lawyers facing court sanctions after filing briefs with case citations an AI invented.
• Intellectual property: creative and design fields face open questions about ownership of AI-generated work and about training data.
• Professional codes: many associations now publish guidance on AI.

You don't need to be an expert in all of this. You do need to know which rules apply in your field, and who to ask.

Try it small, then judge trust

Pick one workflow from your research and try a small version of it yourself, using public or made-up data, never real client, patient or student information.

Test it on several realistic cases, not just one easy example. Include a hard case and one where you already know the right answer, so you can see how it fails.

Then write down three things:

• What worked: where did it save time or catch something you'd miss?
• What didn't: errors, omissions, or places it sounded confident and was wrong.
• What you'd need to trust it: accuracy on your field's cases, a checking step, a tool approved for the data, and clarity about who's accountable.

If you can, talk to a practitioner. Ask what they tried, what stuck and what they dropped, and why. That conversation is often more useful than any demo. Being able to explain all this in an interview shows judgment employers look for.

A public health student maps one workflow

A public health student wants to know how AI is used to communicate research to the public.

Finding real uses: in the video library and in a professional association's newsletter, they find health departments drafting plain-language summaries, researchers using AI to code open-ended survey answers, and a demo of a chatbot answering symptom questions. They count the first two as real uses (named teams, human review) and the chatbot as a demo.

The rules: health information about individuals is protected, so they use only a published, public report.

Trying it: they ask an assistant to summarize the report for a community newsletter at a middle-school reading level. The summary reads well. Checking it against the report, they find one number rounded in a way that changed its meaning and one recommendation the report never made.

What they'd need to trust it: a fact-check against the source every time, a reviewer from the health team, and a tool approved for any non-public data.

They end with a one-page note: three uses, one workflow tested, two errors found, and what would make it safe to use.

Try it

A one-page note on your field: three real uses with where you found them, the rules that apply, and what worked, what didn't and what you'd need to trust the workflow you tried.