boisestate.ai FAQs

boisestate.ai is Boise State’s own generative AI/chat bot model tool, created and supported by university staff. It functions the same as ChatGPT and Gemini, it’s just another option for faculty, staff and students to use. It is especially useful in teaching settings to share conversation results with classes, etc.

More information about boisestate.ai is on our AI at Boise State website.

If our definition of a "closed system" is that all the conversations are in boisestate.ai controlled resources and that none of our interactions with AI are shared or used to train models, boisestate.ai is a closed system.

That said, we still need to apply the data policy rules in boisestate.ai in the same way we would with any other supported generative AI (GenAI) tools.

• Generative Artificial Intelligence (AI) – Guidance on Use and Applicable Policies

• Information Privacy and Data Security (Policy 8060)

Yes, You can share what you create in boisestate.ai with other students, staff, or faculty at Boise State.

Conversations, system instructions, and AI assistants can all be shared with specific Boise State email addresses, using a share link, or made public within the Boise State community. Once shared others will be able to use what you have created to help them with their own tasks.

boisestate.ai is designed an maintained in accordance with Boise State’s Generative Artificial Intelligence (AI) – Guidance on Use and Applicable Policies.

Independent privacy and accessibility statements are in queue and require review in the Office of the General Council.

API access to boisestate.ai release to Students, Staff, and Faculty on Oct 2025.

Refer to Using API Keys for details on using this feature.

AI models can make mistakes for a variety of reasons:

• Ambiguous training data can trigger the AI to guess the most plausible response.

• Outdated or incomplete information or capped training data.

• AI pattern recognition / word prediction can be excellent, but AI often lacks “fact-checking” abilities and this may produce incorrect information.

Always check AI generated content for accuracy. No method for increased accuracy is fool proof and none guarantee accurate responses, but the following may help you more quickly assess the relevancy and accuracy of AI generated responses using the default Claude Sonnet 4 model:

• Consider uploading files, providing direct URLs to the content of interest, and/or selecting “fetch” in conjunction with “brave web search” in advanced settings.

• Be specific in your prompting.

When assessing AI generated content, be aware:

• Some web content cannot be “fetched” due to the web platform, crawl blocking, etc., and this may influence the accuracy or completeness of the AI response.

• If content has been fetched previously, it may have been stored then reused even if it is not the most recent content available, which may mean the response is less relevant.

• Web search may be influenced by URL redirects, archived pages, pages with similar or synonymous content, thus rendering the “meat” of the response worth validating on your own.