Summary and Setup

Understand the opportunities and risks of using AI tools, including large language models, in research conducted on Sagehen HPC. This workshop covers the current AI landscape, Pomona College policies for AI use, data classification considerations, and guidance on whether to use local models on Sagehen or cloud-based AI services. You’ll learn responsible practices for protecting sensitive research data, ensuring research integrity, and using AI tools ethically. This workshop is part of the Pomona College HPC Workshop Series.

Callout

Pomona College HPC Context

This workshop is part of the Pomona College Research Computing Workshop Series for researchers using the Sagehen HPC cluster.

Prerequisites:

  • Completed Workshop 13 (HPC Security Orientation)
  • Understanding of data classification (Workshop 14) recommended
  • Required for all users running AI/ML workloads on Sagehen
  • A Sagehen HPC account (request from its-hpc@pomona.edu)
  • Access to the OnDemand portal

For questions or support, contact the Research Computing team at its-hpc@pomona.edu.

Designed for all researchers and students, this workshop assumes no prior AI knowledge. You will gain practical understanding of how to responsibly leverage AI tools while protecting your research, respecting intellectual property, and adhering to Pomona College policies.

Acknowledgments

Developed by Andrew Wilson, Director of Research Computing and Digital Scholarship at Pomona College, with Andrei Motchenko, who tested, edited and produced screenshots for the workshop series.

Setup: Responsible Use of AI Tools on HPC

Before the Workshop

Software and Accounts

  • Web browser (Chrome, Firefox, Safari, Edge)
  • Optional: Pomona NetID for examples
  • Optional: Access to Sagehen HPC (not required, but helpful for local AI demo)

No Installation Required!

This workshop uses cloud AI assistants and an optional local setup. No special software needs to be installed in advance.

Optional: Try the Tools in Advance

If you want to get familiar with external AI tools, the three most commonly used at Pomona are:

Each offers a no-cost tier. Sign in with a personal account for the workshop exercises — we will only ever use PUBLIC example data in them.

Callout

Product Names Change Faster Than This Lesson

Google’s assistant was called Bard until February 2024, when it was renamed Gemini; bard.google.com now redirects to gemini.google.com. Similarly, chat.openai.com now redirects to chatgpt.com.

Expect names, tiers and model numbers to keep moving. Nothing in this workshop depends on a particular product — the rules you will learn are about the data you are handling, which is exactly why they outlast the tools.

During the Workshop

Tools We’ll Use

External AI (Cloud-Based)

  • ChatGPT, Claude or Gemini for examples
  • A personal account is fine; we use PUBLIC data only
  • Works from any browser

Local AI (Optional, for sensitive data)

  • Sagehen HPC access (request from ITS if needed)
  • Python 3.10+
  • Transformers library (installed during demo)

Software for Hands-On Activities

If You Have Sagehen HPC Access

BASH 

# Log in
ssh your_netid@sagehen.hpc.pomona.edu

# Load conda. There is no pytorch/tensorflow module on Sagehen --
# frameworks live inside your conda environment.
module load miniconda3
conda activate ml_env

If You Don’t Have Sagehen HPC Access

No problem! You can:

  • Observe demonstrations of local model setup
  • Use ChatGPT/Claude for practical examples
  • Learn the concepts without hands-on coding

Workshop Structure

This is primarily a concepts and policy workshop, not a coding workshop.

Part 1: Understanding AI Tools (Classroom)

  • Landscape of AI tools available
  • How AI systems work (high-level)
  • Cloud vs. local AI tradeoffs

Part 2: Data Governance (Classroom)

  • Pomona’s data classification system
  • Regulatory requirements (FERPA, HIPAA, export control)
  • How to classify YOUR data

Part 3: Responsible Practices (Classroom + Optional Hands-On)

  • Bias and fairness in AI
  • Documentation and disclosure
  • Institutional policies and compliance

Part 4: Practical Implementation (Optional Hands-On)

  • Setting up a local model on Sagehen (for RESTRICTED data)
  • Building responsible AI workflows
  • Asking for help from ITS

Key Concepts We’ll Cover

1. Data Classification (Core Skill)

You’ll learn to categorize your research data:

  • PUBLIC: Safe with any AI tool
  • PROPRIETARY: Local AI only or with written permission
  • RESTRICTED: Local AI only + encryption

2. Tool Selection (Core Skill)

Given data classification, choose appropriate tool:

  • ChatGPT/Claude/Gemini? (Only for PUBLIC)
  • A local open-weight model on Sagehen? (For PROPRIETARY and RESTRICTED)

3. Compliance (Essential Knowledge)

Understand:

  • FERPA (student data) - can’t use external AI
  • HIPAA (medical data) - can’t use external AI
  • Export control - research must stay on US servers
  • Pomona policy - requires documentation

4. Responsible Practices (Essential Skills)

  • Verify AI outputs for accuracy
  • Disclose AI usage in publications
  • Audit for bias
  • Document decisions

Before the Workshop: Self-Assessment

Optional: Rate your familiarity with these concepts:

AI Tools:
☐ Never used ChatGPT/Claude
☐ Used once or twice
☐ Regular user
☐ Power user

Data Privacy:
☐ Not familiar with FERPA/HIPAA
☐ Heard of them, unclear on details
☐ Familiar with basics
☐ Work with regulated data regularly

Research Background:
☐ First year graduate student
☐ Mid-career graduate student
☐ Faculty/postdoc
☐ Undergraduate researcher

Your AI Use Case:
☐ Haven't thought about using AI
☐ Considering using AI for research
☐ Already using AI
☐ Not sure if I should be using AI

This helps us pitch at the right level!

Getting Help Before the Workshop

Questions About:

  • Sagehen access:
  • Workshop content: Ask instructor before session
  • Pomona AI policy: Contact your department chair or OVPR

Resources to Review (Optional)

Technical Requirements

Minimum

  • Web browser
  • Internet connection
  • No software installation needed
  • Notepad or note-taking app (for your own data classification notes)
  • Pomona NetID and password (to try local Sagehen examples)

Accessibility

This workshop is designed to be inclusive:

  • Visual aids and text slides
  • Spoken descriptions of code examples
  • Discussions don’t require technical background
  • Hands-on activities are optional

Let instructors know if you need accommodations!

Contact for Setup Issues

Before the workshop, if you encounter any issues:

  • Sagehen access problems:
  • Browser/tech issues: Your local IT support
  • Workshop-specific: Email instructor

Post-Workshop Resources

After the workshop, continue learning:

These will be provided at the end of the workshop.