← → or click edges to navigate · F for fullscreen
Drexel UniversityDornsife School of Public Health
Department of Environmental
& Occupational Health
AI WORKSHOPS FOR DSPH

Foundations & Capabilities

What generative AI is, which Drexel tools to use, and how to prompt for useful results

Alex Quistberg, PhD, MPHAssociate Research Professor · Department of Environmental & Occupational Health
Workshop 1 · June 29, 2026
WHAT YOU'LL LEAVE WITH

Agenda & Objectives

  1. 1Explain what a language model can and can’t do.
  2. 2Pick the right Drexel tool and follow the data rules.
  3. 3Sign in to ChatGPT Edu and run useful prompts.
  4. 4Use a simple framework (PTCF) for better outputs.
  5. 5Recognize how AI fails and who is accountable.
Environmental & Occupational HealthDornsife School of Public Health
SECTION 01

What AI Is — and Isn’t

How large language models work, what they are good at, and where they go wrong.

THE MENTAL MODEL

AI Drafts, You Decide

Tools like ChatGPT and Copilot are large language models: they predict likely text from patterns in huge amounts of writing. They are strong at drafting, summarizing, rewriting, explaining, and brainstorming.

  • They don’t “know” facts like a database — and can state wrong things confidently (a “hallucination”).
  • Treat AI like a fast, well-read assistant who is sometimes wrong and never offended by edits.
  • You stay in the loop: AI produces a draft; you verify and decide.
Environmental & Occupational HealthDornsife School of Public Health
TOOLS & DATA RULES

Which Tool — and the Data Rule

Approved for Drexel work
  • ChatGPT Edu (use GPT-5.5 when available)
  • Copilot Chat with Enterprise Data Protection; Microsoft 365 Copilot
  • Zoom AI Companion and Adobe tools, per Drexel’s data classifications
Low Risk Data only
  • Claude, personal ChatGPT, Gemini, Perplexity — unless separately approved
  • Never paste PHI, sensitive PII, or protected records into an unapproved tool
  • When in doubt, check the Tools & Access page before you paste
Environmental & Occupational HealthDornsife School of Public Health
SECTION 02

Ethics & Governance

How AI fails, and who is accountable — the thread that runs through every workshop.

HOW AI FAILS

Two Failure Modes, One Rule

Confident Errors

Models predict plausible text, so they can invent facts, statistics, and citations that look right and are wrong.

Bias

Models learn from internet-scale data that carries society’s inequities, so outputs can reflect and amplify stereotypes.

Who’s Accountable

The person who acts on AI output owns the result — not the tool, not the vendor. AI drafts; a human verifies and decides.

Environmental & Occupational HealthDornsife School of Public Health
WHY IT MATTERS IN PUBLIC HEALTH

Bias Has Real Consequences

A widely used US healthcare algorithm used past healthcare cost as a proxy for health need. Because less had historically been spent on Black patients, it rated them as healthier than equally sick White patients (Obermeyer et al., 2019).

  • Bias enters through training data, representation, and proxies that don’t mean the same thing across groups.
  • The lesson isn’t “don’t use AI” — it’s to check outputs, especially across groups.
Environmental & Occupational HealthDornsife School of Public Health
SECTION 03

Hands-On

Sign in, run your first prompts, and learn a simple recipe for better results.

ACTIVITY 1 · ~17 MIN

First Prompts in ChatGPT Edu

Sign in to ChatGPT Edu with your Drexel account. Use public or made-up details only — no sensitive data. Run each prompt and notice what it does well.

  • Summarize — turn a public report into 5 bullets and a one-line bottom line.
  • Draft — a friendly, professional reminder email under 120 words.
  • Explain — define a “data use agreement” in plain language, about 100 words.
  • Reformat — turn rough notes into a clean numbered agenda with time estimates.
Environmental & Occupational HealthDornsife School of Public Health
ACTIVITY 2 · ~15 MIN

Better Prompts with PTCF

A little structure turns a generic answer into a usable one. Build your prompt from four parts:

  • Persona — who should it act as? “You are an experienced grants administrator…”
  • Task — what exactly should it do? “Draft a 150-word reminder…”
  • Context — what background does it need? “…for faculty submitting NIH progress reports.”
  • Format — how should the output look? “Bulleted, warm tone, deadline at the top.”
Environmental & Occupational HealthDornsife School of Public Health
WRAP-UP

What to Take Away

AI is a drafting and thinking assistant — you verify and decide.

  • For Drexel work: ChatGPT Edu or Copilot, and mind the data rules.
  • PTCF (Persona, Task, Context, Format) is your default recipe for a good prompt.
  • Before Workshop 2: do one weekly task in ChatGPT Edu using PTCF, and bring it.
Environmental & Occupational HealthDornsife School of Public Health
Drexel UniversityDornsife School of Public Health
THANK YOU

Questions & Discussion

Next: Workshop 2 — Workflow Enhancement. Use Copilot and ChatGPT Edu to speed up the work you already do.

Alex Quistberg, PhD, MPHdaq26@drexel.edu
Department of Environmental & Occupational Health