19 Working with AI

Our notes on working responsibly and effectively with AI coding assistants have moved to a dedicated site: Working with AI (source: https://github.com/Morrison-Lab/wai).

That site covers:

Policies for Using AI: responsibility for validation, disclosure, attribution, and using AI for journal articles
Coding Agents: what language models, coding agents, and harnesses are; how to work with them; benefits, hazards, and best practices; and configuring your environment
Pull-Request Workflow with Agents: filing issues, claiming work, and driving a pull request to a clean, mergeable state

Lab members who use AI tools must still adhere to the guidelines described there.

Looking for a specific section that used to be here?

This chapter previously had section anchors like #sec-ai-best-practices or #sec-ai-disclosure for each subtopic. Those anchors still resolve on this page (so old bookmarks and links won’t 404), but the content itself now lives on the linked wai pages above — follow the matching bullet to find it.

19.1 Institutional AI account access

19.1.1 UC Davis Health OpenAI access

For lab members affiliated with UC Davis Health (UCDH) or conducting health sciences research, enterprise OpenAI accounts are provisioned through the UCDH IT Self-Service (ITSS) ServiceNow catalog.

To request an account:

Navigate to the UCDH ServiceNow OpenAI catalog item.
Log in with your UC Davis Health credentials and complete Duo two-factor authentication if prompted.
Fill out the request form with your department, contact details, and business or research justification.
Submit the request for processing by UCDH IT.
Once approved, follow the instructions in the confirmation email to join the UC Davis Health OpenAI workspace.

19.1.2 UC Davis campus OpenAI access

For UC Davis campus affiliates, access to ChatGPT Edu is managed through Information and Educational Technology (IET) Aggie AI. See the IET Aggie AI ChatGPT Edu documentation for setup instructions and eligibility details.

19.1.3 Data security and compliance

When using institutional AI accounts:

Never submit Protected Health Information (PHI), personally identifiable information (PII), or restricted research data to AI tools unless authorized under an approved protocol.
Follow all institutional data handling policies and the lab’s guidelines on Working with AI.