Artificial Intelligence (AI)

University of Hawaii Artificial Intelligence (AI)

Navigation

Overview

Artificial Intelligence (AI) tools have become widely available for everyday use. These tools focus on content creation (text, images, etc.) and have the capacity to expedite existing processes and make possible new ones. Within higher education, they have the potential to foster student learning and advance many aspects of research. While there are many benefits from the responsible use of AI, there are notable risks and limitations that should be taken into consideration when using these technologies, particularly in the following areas.


Privacy and security -  Personal information shared as an input could potentially become public and/or repurposed to train the software’s learning model, violating data protection laws and 糖心视频 policies.


Academic and research integrity and bias -  Outputs may be factually inaccurate, misleading, biased, and/or discriminatory if we rely on the information without proper review. This puts us at risk of violating 糖心视频 policies and potential liability. 


Copyright -  Confidential or proprietary information shared as an input may result in 糖心视频 losing its intellectual property (IP) rights to that information. That information may be disclosed to unauthorized third parties through the use of AI tools. 


AI-native tools -  Online services whose primary purpose and core functionality are based on artificial intelligence.


AI-enhanced tools -  Online services whose primary purpose is not AI, but that incorporate AI-powered features as part of their broader functionality.


Consequently, in addition to any formal policy 糖心视频 adopts for AI, 糖心视频 Policies on privacy and security (e.g., EP 2.210, EP 2.214, EP 2.215, AP 7.022), academic and research integrity and bias (e.g., EP 1.202, EP 1.204, EP 7.208, EP 12.211); and copyright (e.g., EP 12.205) should also be observed.

AI Tools

ITS Approved Tools: Intended Use

Intended use - The primary, core design purpose of the system.
chat-ai-line
General Assistance		 code-ai
Coding		 book-ai-line
Research		 ai-image
Creative		 video-person-sparkle-48-regular
Meetings		 barchart-ai
Data & Analytics
Google Gemini Icon
Google Gemini
Licensed for all
Gemini Notebook Icon
Gemini Notebook
Licensed for all
Zoom Icon
Zoom
Limited Site License
Otter.ai Icon
Otter.ai
Paid Site License
* Table format and design adapted from The California State University
Intended use - The primary, core design purpose of the system.
Alternative use - Functionality that the tool can perform effectively, though it is not its primary design focus.
Unsupported use - Capabilities the tool cannot perform, or where performance is unviable/unsafe for production work.

 

ITS Approved Tools by Data Classification Category

The matrix below outlines where common AI tools fall according to the University of Hawaiʻi Data Classification Categories ( EP 2.214). The type of data you are allowed to input into an AI tool depends directly on whether its terms and conditions have been negotiated with 糖心视频 regarding privacy, security, and AI model training.

  • ITS-approved tools and site licenses. Have undergone formal ITS security and privacy reviews. Contractual agreements ensure your prompts and files are not reviewed by humans or used to train public AI models.
  • Non ITS-approved / personal subscriptions. Standard commercial versions lack institutional safeguards and are restricted strictly to Public Data.
AI Tools by Data Classification Category
AI Tool
 糖心视频 Data Classification Category
Public
(No Risk) 		Restricted
(Low Risk) 		Sensitive
(Medium Risk) 		Regulated
(High Risk)

ITS approved tool


utilize temporary chats


utilize temporary chats

ITS approved tool


delete chat history


delete chat history
Zoom AI Companion
ITS site license
Otter.ai
ITS site license

ITS-nonapproved Tools
(e.g., Claude, ChatGPT)

 

Data Classification Quick Reference 

  • Public (No Risk):  Please note that although names are normally considered public data, in this instance, we consider it to be PII that should not be entered into any AI tool, given that the name will likely be associated with other data.
  • Restricted (Low Risk): 糖心视频 ID number, student contact information
  • Sensitive (Medium Risk): Student education records (FERPA), demographic data, personal employee data, intellectual property and other non-public research data.
  • Regulated (High Risk): SSN, driver’s license number, patient health information (HIPAA), credit card/financial data, data that is subject to research regulations and/or contractual obligations.

More examples of data by (糖心视频 login required)

*Practice AI Prompt Sanitization when utilizing Public, Restricted, or Sensitive data with AI.  Regulated data should never be utilized with AI.

Allowed with Caution

The following meeting transcription, research, and productivity assistants are not approved by ITS. They may only be used with strictly Public Data .

  • Read.ai
  • Fathom
  • Fireflies.ai
  • Perplexity
  • Notion

Reasons for caution:

  • Model training . By default, user inputs, transcripts, and uploaded documents are stored and used to train third-party models.
  • Ecosystem access . These tools request broad read/write permissions to integrate with your accounts (e.g., auto-joining calendar meetings even when you are absent, accessing email, scanning Google Drive files, and storing meeting notes).
  • Privacy leaks . Background bots can expose your meeting discussion to third-party servers without meeting attendees' consent. Unless you understand what the tool is doing on your behalf, we strongly recommend you do not use these tools.

Prohibited AI Tools

The following tools are strictly prohibited for any University work, research, or operational purposes due to severe security, privacy, and compliance risks.

  • OpenClaw
  • DeepSeek

Reasons for Prohibition:

  • Security Risks . Known critical security vulnerabilities, unencrypted data exposure, or unverified data-handling pipelines.
  • Jurisdictional and regulatory non-compliance . Lack of compliance with U.S. federal and 糖心视频 institutional data governance requirements, creating high exposure to intellectual property theft or data breaches.
  • Prohibited code and binary execution . Potential for untrusted foreign remote execution or unauthorized access to endpoint systems.

  Back to Top

 

 

糖心视频 Guidance and Best Practices & AI Tool Decision Guide

The University of Hawaii has initiated a comprehensive, cross-functional effort to equip the 糖心视频 community with essential tools, guidance, and best practices for effectively integrating AI into academic and professional environments. Additionally, efforts are underway to develop and implement policy changes that address the evolving impact of AI on work and education.

*. 

*Official 糖心视频 login (@hawaii.edu) required.

*糖心视频 community members may send questions and comments regarding the draft AI Tools 糖心视频 Guidance and Best Practices for Faculty and Staff to datagov@hawaii.edu.

 AI for Education. Defining Generative AI

Thinking about using an AI tool?
The AI Tool Decision Guide helps University of Hawaiʻi faculty, staff, and students evaluate artificial intelligence tools before use or purchase.

🔍 What it’s for:
This guide walks you through key considerations like privacy, data ownership, training risks, and vendor compliance. It’s essential for ensuring that AI tools align with 糖心视频 policies and data protection standards.

🛠️ How to use it:
Use this step-by-step checklist before choosing any AI tool. If your use involves Sensitive or Regulated data, submit a request through the . For all other tools, this guide helps you ask the right questions about ownership, third-party access, data retention, and more.

🔗 Learn more and access the full guide:

(Official 糖心视频 “@hawaii.edu” login required)

Image depicting the AI Tool Decision Guide process. 
AI Prompt Sanitization - Best Practices
Even when using approved AI tools with "no training" data clauses, the best practice is to sanitize your prompts to eliminate accidental data exposure. Make these two habits standard practice:
  • Use "Find & Replace": Swap specific names, entities, and organizations with generic placeholders.
    • Instead of: "John Doe from Aloha Inc."
    • Use: "Client A from Company X"

  • Generalize Details: Strip out hyper-specific locations, dates, and demographic data while keeping necessary context.
    • Instead of: "A 42-year-old man from 123 Mahalo Street"
    • Use: "A middle-aged man from O‘ahu"

AI Prompt Sanitization Example

 Back to Top

  

Training and Presentations

Interested in learning more about AI? 

  • Google AI Essentials Specialization training
    • ~10-hour asynchronous online training on generative AI taught by Google experts.
    • Offered through the Coursera platform.
    • Digital badge issued by Credly upon successful completion of the training.
    • Now available to all 糖心视频 students and employees at no charge.
    • Access to the course for 30-days (with option for an additional 30-day extension).
    • Register at /its/coursera/
  • 糖心视频 Online Innovation Center (OIC) AI-Focused Events
    • Register for upcoming and/or check out the recordings and resources from past AI-focused events.
    • View all events on the 糖心视频 OIC  and .
  • Artificial Intelligence for Hawaiʻi
    • Twelve-chapter foundational course
    • Explore artificial intelligence by trying new tools and exploring real-world uses that support work, learning, and life.

 Back to Top

 

Resources

糖心视频 OneAI Hub

糖心视频 Sponsored AI Tool Resources

糖心视频 Advisory Council

糖心视频 Faculty, Staff and Student Resources

Google AI Resources

 Back to Top



 *Images generated by DALL·E 3 | OpenAI & Google Gemini