Tool Overview and User Eligibility

Where can I learn about PhoenixAI and its features?

Visit the University's generative AI website for guidance, resources, and training.

How do I access PhoenixAI?

Visit the PhoenixAI website and log in with your UChicago CNetID and password.

What is PhoenixAI, and how does it differ from external AI tools?

PhoenixAI is a generative Artificial Intelligence (AI) tool, which is built on open-source technologies and Large Language Models (LLMs), and tailored specifically for the University of Chicago community.

The advantages of PhoenixAI over external tools include:

• Data Security and Privacy: Your data (queries, model responses, documents you upload, interaction logs, etc.) is protected within a University-owned environment and cannot be used by any vendor. The environment is only accessible via UChicago single-sign-on and multi-factor authentication.
• Equitable Access: It provides access to premium models that usually charge monthly fees for each user.
• Accessibility: The interface is built with accessibility features not always available in commercial offerings.

What are the main features of PhoenixAI?

Some of the main features of PhoenixAI include:

• Chat Interface: Similar to ChatGPT, PhoenixAI allows for dynamic interaction with queries and responses. It is an accessible interface, built to the University of Chicago Digital Accessibility Standards.
• Data Security: Data submitted to PhoenixAI is kept within the UChicago environment and not shared with third-party vendors or used to train AI models.
• Prompt Templates: Users can customize the chatbot's behavior and characteristics by using a prompt template. To access the prompt library, visit the university's generative AI website.
• Feedback Options: Users can provide feedback on responses by using the thumbs up (good response) or thumbs down (bad response) icons located at the end of each prompt response.

What does it mean that PhoenixAI is a beta version?

The University plans to continue adding new functionality to PhoenixAI to better fit the needs of the University community. While using the beta version, you may experience issues with the responsiveness and performance of the tool.

Can we create Assistants (Formerly Custom GPTs- Generative Pre-trained Transformers) with this beta version?

Assistants are not the same as Large Language Model (LLM) apps that use APIs. With the November 2024 release, users can create and share their own Assistants. Visit PhoenixAI Assistants: Create, Edit, Share, Delete a Assistant for step by step instructions and resources about Assistants. Please note that Assistants are only available with the GPT-4.1 model.

What is the current GPT model being used for PhoenixAI?

As of June 18, 2025, PhoenixAI users have the option to choose between two current models: GPT-4.1, and o3. To learn more about the new o3 model and how it differs from GPT-4.1, read our PhoenixAI GPT Model Differences article.

Who is eligible to use and access PhoenixAI at the University of Chicago?

UChicago faculty, other academic appointees, postdoctoral researchers, students, and staff can access the beta version of PhoenixAI by logging in with their CNetID and password. This includes members of the Biological Sciences Division.

UChicago Medicine (UCM) employees who do not have CNetIDs are not eligible to access the PhoenixAI tool as it does not meet UCM data security and privacy requirements.

Note: Restricted data such as protected health information (PHI) and research health information (RHI) should not be used in PhoenixAI or other generative AI tools.

Is PhoenixAI accessible to users with disabilities?

PhoenixAI has been built to conform with the University of Chicago Digital Accessibility Standards. To provide feedback about the accessibility of the tool, please contact ai-feedback@uchicago.edu. To report an accessibility barrier, please use the Report a Digital Accessibility Barrier form.

I am looking to learn more about the usage of PhoenixAI and its different functionalities.

Please refer to the following LinkedIn Learning course to receive an helpful suggestions on how to write prompts and use GenerativeAI: How to Boost your Productivity with AI Tools. To view additional resources, visit the University's generative AI website.

Are there any limitations to using PhoenixAI?

• In the beta version of the tool, you are limited to 50 queries per day. You will receive a message once you have hit this quota. This limit resets at midnight CST.
  • For the o3 model, the limit is 50 queries per day for faculty and 10 queries per day for all other users.
  • The 50 prompt limit is specific to the GPT-4.1 model for both faculty, students, and staff.
• Because PhoenixAI is built based on the current Open AI model, it cannot properly, if at all, respond to prompts asking about events that occurred after June 2024.
• All GenAI tools can give a variety of responses based on your prompts. PhoenixAI can make mistakes, so cross reference the responses as appropriate.
• PhoenixAI has not been trained on UChicago data sources or facts. Its knowledge about the University is only based on publicly available information that was included in its training data.
• As it is currently designed, PhoenixAI will not learn as more UChicago users interact with it because it is not being trained on this data. Its capabilities will change and improve only if the underlying model gets upgraded.
• The o3 model is experimental in nature. Sensitive / confidential data, such as PHI and PII, should not be used when entering a prompt.
• A user cannot switch between the models GPT-4.1 and o3 in the same chat / thread. Only one model can be utilized per chat / thread.
• Due to the complex reasoning power of the o3 model, response will take longer compared to the GPT-4.1 model.

All of these limitations are subject to change as we develop our roadmap based on input from the University community.

Data Privacy and Ethics

What kind of data security measures are in place for PhoenixAI?

Data submitted to PhoenixAI is kept within the UChicago environment and not shared with third-party vendors or used to train AI models. Please see Policy 601 for definitions of confidential data and its use for more information.

Note: Restricted data should not be used in PhoenixAI or other generative AI tools.

How is the privacy of my interactions with PhoenixAI ensured?

Data submitted to PhoenixAI is kept within the UChicago environment and not shared with third-party vendors or used to train AI models. You can opt to have PhoenixAI retain your prompt data for subsequent sessions or delete the prompts that you do not want the tool to store after the session is completed.

What is an Incognito Chat?

To expand PhoenixAI's privacy protocols, Incognito Chats add a layer of security by automatically deleting all files, prompts, and responses 30 minutes after creation. These chats are private between the user and PhoenixAI only. During an Incognito Chat, users cannot access previous threads or share the current thread. For security purposes, closing the PhoenixAI tab or starting a new chat instantaneously deletes the Incognito Chat. Once an Incognito Chat has been closed or it reaches the 30-minute timer, there is no way to retrieve the information discussed in the thread. Additionally, Incognito Chats do not show up in the History section.

How do I create an Incognito Chat?

Review the steps listed within Create and Use Incognito Chats in PhoenixAI.

How does PhoenixAI handle sensitive or confidential information?

PhoenixAI is built in secure and private environment that protects University data from being used by third parties for any purpose.

Can I share the outputs generated by PhoenixAI publicly?

This is a general-purpose tool we expect users to employ in many different ways. You are welcome to share the outputs as long as you are following your data security obligations and ethical standards for the use of generative AI.

Effective Use Practices

Are there any guidelines or best practices for using PhoenixAI effectively?

View the University's generative AI guidance. For more information, visit the resources on the University's generative AI website.

How can faculty and researchers leverage PhoenixAI for academic purposes?

PhoenixAI was created to be used broadly and creatively by the UChicago community. Visit the University's generative AI website for resources, guidance, and training on using generative AI in teaching and learning.

As a researcher, make sure that your use of PhoenixAI Beta does not violate any agreements you have made in the Research Data Protection Policy.

• If your research has been reviewed by IRB, ask your IRB if you may use PhoenixAI Beta.
• If your research involves a data use agreement (DUA), find out if your DUA allows the use of generative AI tools.
• If neither of these apply, please contact ai-feedback@uchicago.edu.

The o3 model may be best suited for academic purposes as it prioritizes response accuracy for advanced purposes. To learn more about the new o3 model and how it differs from GPT-4.1, read our PhoenixAI GPT Model Differences article. 

How can staff and faculty leverage PhoenixAI for research using University institutional data?

Performing research on University institutional data (e.g. proprietary research data, student information, unpublished academic work), requires several steps to make sure that the University's privacy interests are upheld. Details are explained on the Data Stewardship Council's website. Contact ciso@uchicago.edu to start the process. Once you've secured permission to use the data, you can use it on PhoenixAI.

How can I customize the way that PhoenixAI responds to prompts?

PhoenixAI provides a way to customize prompts to better fit your needs.

Follow the steps to customize prompts:

1. Select the user icon.
2. Click Settings to open the Settings dialog box.
3. Choose Customize Chat from the General option on the left.
   
4. Input the custom prompt template.
5. Click Save to return to the Settings options.
6. Select the X in the top right corner to dismiss the Settings box.

Customize Chat Prompt Template Example:

For example, if you want to see all times in Chicago time using the 12 hour format rather than 24 hour format (commonly used in the military), you can specify this by entering "Dates and times should be displayed in Central Standard Time, in 12 hrs, not 24 hrs."

How can I clean up my prompt/chat/thread history? What options do I have?

PhoenixAI allows several options for managing your threads (also referred to as chats or prompts). Hover over the thread to see the Options menu, represented by three vertical dots.

You are able to share, rename, archive or delete any thread in your History. 

I accidentally deleted a prompt. Can I get it back?

Unfortunately, once a prompt is deleted from the sidebar, it cannot be recovered.

How does PhoenixAI manage access controls for generated files?

• Only an authenticated user can download a file that PhoenixAI has generated for them.
• The link to a file that PhoenixAI has generated will not be present moving forward as it will be a button now for the user to click on to the download the file.
• Previous files that were generated by PhoenixAI will not be downloadable to the authorized user. Users will need to recreate the file.

Why don't sources always appear for internet searches?

While we aim to provide sources whenever possible, they may not always appear due to technical limitations. Sometimes sources are available and linked, other times, they may not be directly cited. If you need verified sources for critical information, we recommend cross-checking with trusted websites.

Future Technical Enhancements

Will PhoenixAI offer access to other Large Language Models (LLMs)?

In future versions of the PhoenixAI tool, we plan to offer access to additional LLMs. Currently the tool is built on OpenAI's GPT-4.1 and o3 models.

How can I use PhoenixAI for my own AI projects?

Please contact ai-feedback@uchicago.edu if you are interested in using PhoenixAI technology to support University work and research.

Can I integrate PhoenixAI into existing applications or platforms?

PhoenixAI is intended as an interactive stand-alone product for the UChicago community. If you intend to integrate generative AI features into your applications or workflows, we recommend you consider using LLMs via the API method. Please contact ai-feedback@uchicago.edu to discuss your options.

Troubleshooting Guide for Common Issues

This troubleshooting guide provides instructions to help you diagnose and resolve common issues you may encounter while using the PhoenixAI tool.

The screen takes a long time to load.

Refresh your browser.

The screen is blank and I've used PhoenixAI before.

Follow these steps to attempt to narrow down the issue if you've previously been able to get into PhoenixAI:

1. Verify you have an internet connection and are able to access other University sites like https://workday.uchicago.edu or https://canvas.uchicago.edu
2. Completely close out of your internet browser and try a new session.
3. Try connecting using a private or incognito browser window. If you can access PhoenixAI using this method, try clearing your browser's cache and cookies. Then try again connecting in your usual browser (non-private) window.
4. If you still can't connect, please submit a request to the IT Services Self-Service Portal or call 773.702.5800 for immediate assistance.

The tool froze and isn't providing a response after I immediately sent a prompt upon logging in.

In this case, the tool hasn't fully loaded, so follow these steps to reboot.

1. Refresh your browser.
2. Wait for your chat history to load on the sidebar.
3. Re-type and submit your prompt.

I want to upload a file but don't see any option to do so.

Select the paperclip icon on the message bar to upload files. PhoenixAI can respond to questions about data in the uploaded file. Acceptable file types include: c, cpp, css, csv, docx, html, java, js, json, md, pdf, php, pptx, py, rb, tar, tex, ts, txt, xlsx, xml, zip.

Note: Only the GPT-4.1 model has the ability to upload files. 

Error: File size is too large

If you upload a file larger than 20MB you may receive a "File Size Is Too Large" error. In that case, you will have to use a smaller file. If you receive the same error but the file is under the 20MB limit, this may be due to the number of words present within the file. To circumvent this issue, create a new Assistant and you can upload the file to perform any analysis that you want.

This solves the issue because PhoenixAI interprets the textual content of files within fixed size chunks or tokens. Word length determines the number of tokens needed so although the main PhoenixAI thread can handle up to 40,000 tokens, this roughly translates to 32,000 words. However, Assistant threads do not have the same token limit as the main PhoenixAI thread.

Otherwise, if you receive an error message after uploading an acceptable file type, such as "page not found" or "500 - internal server error," you may want to try again. If the second attempt is unsuccessful, please submit a request to the IT Services Self-Service Portal or call 773.702.5800 for immediate assistance.

Error: Empty file 

When uploading .txt files to PhoenixAI, users may encounter an "empty file" error if the file contains special characters. PhoenixAI currently only supports files made up of plain alphanumeric characters — that is, letters A–Z, a–z, and numbers 0–9. Any characters outside this range (including punctuation, symbols, bullet points, spaces, and control characters) may cause the system to reject the file or return it as empty. 

To ensure successful processing, please remove all non-alphanumeric characters before uploading, or save the .txt file as a Word document and reupload to PhoenixAI. 

How do I use the image analysis functionality?

Upload your image using the paperclip icon on the message bar and specify to analyze. A description of the image will then be provided. For more detailed image analysis, PhoenixAI will recommend using Microsoft Copilot. 

Notes:

• Image analysis is only available in the GPT-4.1 model.
• Video files are not supported.
• A maximum of one file should be uploaded at a time.
• The file should not exceed 20 MB. 

Can PhoenixAI generate images?

Image generation is currently restricted to creating graphs and charts generated from data uploaded to PhoenixAI. Any image generated is immediately available to download.

Example: 

We uploaded an Excel spreadsheet with numbers for three months of tickets resolved by three separate teams. We asked PhoenixAI to analyze the data and create a visual representation of the data. It responded with a line graph, but we asked PhoenixAI to change the representation from a line graph to a bar chart.

Support and Feedback

What types of support or resources are available for PhoenixAI users?

The University's generative AI website is a central hub for information on generative AI tools, including PhoenixAI, guidance, and training. It includes specific resources for faculty, researchers, staff, and students.

How do I report issues or provide feedback about PhoenixAI?

For support with PhoenixAI, submit a request to the IT Services Self-Service Portal or call 773.702.5800 for immediate assistance.

To report an accessibility barrier, please use the Report a Digital Accessibility Barrier form.

If you have questions about future features or want to integrate PhoenixAI into other systems, please contact ai-feedback@uchicago.edu.

Resources

• Faculty can find out more about acceptable uses of PhoenixAI for both themselves and their students on the How to Think About Generative AI in Your Teaching webpage.
• Find out more about Assistants on OpenAI's Introducing GPTs webpage.
• LinkedIn Learning course to receive an helpful suggestions on how to write prompts and use GenerativeAI: How to Boost your Productivity with AI Tools.