UChicago Generative AI Tools
PhoenixAI
PhoenixAI is the University’s generative AI chat service, which is built with the latest large language models (LLM) from OpenAI and customized for the University community for its security, privacy, accessibility, and equity. For more information about its latest features, view the release notes.
To get started, visit phoenixai.uchicago.edu and log in with your CNetID and password.
Features of the beta version include:
-
Assistants: Build custom AI assistants tailored to specific needs. Assistants are designed to handle unstructured data and text-based inputs.
-
Chat Interface: Allows for dynamic interaction with queries and responses.
-
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: Customize the chatbot's behavior and characteristics by using a prompt template. See prompt examples.
PhoenixAI Resources
If you are interested in integrating PhoenixAI technology with other systems at the University, please contact ai-feedback@uchicago.edu.
Microsoft Copilot
Microsoft Copilot is a generative AI-powered chat tool that is integrated with Microsoft’s Bing search engine to offer internet-connected AI answers and image generation. It is based on OpenAI’s GPT-4 and DALL-E 2 models.
Copilot is available to the public as a free tool. However, the University’s Microsoft 365 licensing agreement provides data and privacy protections that the public version does not offer. It is strongly recommended that users log in with their UChicago email while using Copilot for University-related searches. To get started, visit copilot.cloud.microsoft and log in with your UChicago account.
For more information about Microsoft Copilot, view the Knowledge Base article, Microsoft Copilot Overview.
GitHub Copilot Business
The University offers access to GitHub Copilot Business, an AI-powered software coding assistant developed by GitHub, OpenAI, and Microsoft. As a response to the interest in AI applications with code generation capabilities, the GitHub Copilot Business agreement provides this service and provides data protection through the University’s Microsoft contract.
Users of the tool will incur a cost for it and must be part of an existing enterprise GitHub organization. To request this feature, unit leaders may contact IT Services.
Product Comparison
| PhoenixAI | Google Gemini | Microsoft Copilot | |
|---|---|---|---|
| AI Privacy | Data remains in UChicago environment, not used to train models | Data shared with Google for training models | Data not used to train Microsoft models |
| Data Security* | Internal Data or Public Data | Public Data only | Internal Data or Public Data |
| Saved Chats | User can choose | Yes | Yes |
| Image Generation | Not in current version | Yes | Yes |
| Internet Integration | Not in current version | Yes (Google) | Yes (Bing) |
| UChicago Single Sign-On | Yes | No, personal account required | Yes |
| AI Models | GPT-4.1, o3-mini | Gemini | GPT-4, GPT-4 Turbo, DALL-E 3 |
*Restricted data should not be used in PhoenixAI or other generative AI tools.
Approved, Restricted, and Unauthorized AI Tools at UChicago
| AI Tool | Status | Purpose/Reason | Enterprise Supported/Pay/Free | Restrictions | Reviewed by Date |
|---|---|---|---|---|---|
| BoxAI | Approved up to SRDS high protection level. | General Use | Enterprise Supported | Can be used for sensitive information with IRB approval. | January 17, 2025 |
| Copilot | Approved up to SRDS high protection level. | Various purposes to support Microsoft Products | Enterprise Supported | Can be used for sensitive information with IRB approval. | January 17, 2025 |
| Sonix AI | Approved up to SRDS low protection level. | Transcription service | Paid | Only for non-sensitive information. | January 17, 2025 |
| claude.ai | Approved up to SRDS low protection level. | General Use | Paid | Only for non-sensitive information; not to create website chatbots. | January 17, 2025 |
| ChatGPT 3.5 | Approved for data that is made publicly available by its source. | General Use | Free | Only for non-sensitive information | January 17, 2025 |
| ChatGPT 4.0 | Approved for data that is made publicly available by its source. | General Use | Free | Only for non-sensitive information | January 17, 2025 |
| Azure AI | Approved up to SRDS low protection level. | General Use | Enterprise-supported | Only for non-sensitive information | January 17, 2025 |
| ExtraHop AI | Approved for all data types | For information security | Paid | Can be used with sensitive flow data | January 17, 2025 |
| Vertex AI | Approved for data that is made publicly available by its source. | General Use | Paid | Only for non-central GCP tenant | January 17, 2025 |
| 3 Play Media | Approved up to SRDS moderate protection level. | Transcription service, Accessibility | Paid | No SRDS high protection level level data use approval. | January 17, 2025 |
| Zoom AI Companion | Approved up to SRDS high protection level. | Taking meeting notes | Enterprise-supported | Can be used to take notes that include sensitive information | January 17, 2025 |
| Parlay Ideas | Approved up to SRDS low protection level. | Creating class discussions | Paid | Only for non-sensitive information | January 17, 2025 |
| Tunnl Platform | Approved up to SRDS low protection level. |
Allows users to build audiences and helps market to those audiences |
Paid | Only for non-sensitive information |
January 17, 2025 |
| ClassBuddy | Approved up to SRDS low protection level. | Teaching Assistant that ingests uploaded course material from Canvas into its knowledgebase. | Paid | Only for non-sensitive information | January 17, 2025 |
| Site Improve Intelligence | Approved up to SRDS low protection level. | Helps enterprises plan, create and optimize content. |
Paid | Only for non-sensitive information |
January 17, 2025 |
| GatherContent by Bynder | Approved up to SRDS low protection level. |
Helps organizations scale up content production and deliver better quality content faster. | Paid | Only for non-sensitive information |
January 17, 2025 |
| Keen Platform | Approved up to SRDS low protection level. | Marketing decision support | Paid | Only for non-sensitive information |
January 17, 2025 |
| Chat2Learn | Approved up to SRDS moderate protection level. |
Provide conversation topics for parents to talk to their kids. | Paid | No SRDS high protection level level data use approval |
January 17, 2025 |
| Point and Click AI | Approved up to SRDS low protection level. |
Enhance features and efficiency of wellness services provided to students. |
Paid | Only for non-sensitive information |
June 30, 2025 |
| Airtable | Approved up to SRDS low protection level. |
Airtable is a database tool - providing low-code applications for end users. |
Paid | Only for non-sensitive information | June 30, 2025 |
| CampusESP: Family Communication Module | Approved up to SRDS low protection level. |
Helps higher education institutions share content with families across web, mobile, email, and SMS. |
Paid | Only for non-sensitive information |
June 30, 2025 |
| SprinklrAI+ | Approved up to SRDS low protection level. |
Designed to enhance customer experience management across enterprises |
Paid | Only for non-sensitive information |
June 30, 2025 |
| Noodle Partners, PBC: Manage and Engage AI |
Approved up to SRDS low protection level. |
Manage AI: An AI-driven platform that helps improve programs and support students by turning data from systems like LMS and CRM. Engage AI: Uses public program data to engage prospective students in personalized conversations. |
Paid | Only for non-sensitive information |
June 30, 2025 |
| Asana | Approved up to SRDS low protection level. |
Project Management | Paid | Only for non-sensitive information |
June 30, 2025 |
| Xapien | Approved up to SRDS low protection level. |
Research engine used for due diligence on donors and partners. |
Paid | Only for non-sensitive information |
June 30, 2025 |
| Chatfuel | Approved up to SRDS low protection level. |
Administers surveys on Facebook as part of an ongoing research project |
Paid | Only for non-sensitive information |
June 30, 2025 |
| Mail Maestro | Approved up to SRDS low protection level. |
AI email assistant | Enterprise-supported | Only for non-sensitive information |
June 30, 2025 |
| Scribe AI | Approved up to SRDS low protection level. | AI email assistant | Paid | Only for non-sensitive information | June 30, 2025 |
| NotebookLM | Contact us to discuss your specific use case. | An AI -powered research and note-taking tool. |
Enterprise-supported | Contact us to discuss your specific use case. |
June 30, 2025 |
| Gemini | Approved up to SRDS low protection level. | Google’s family of generative AI models (formerly Bard) that provides advanced capabilities for text generation, summarization, reasoning, coding, and multimodal input. | Enterprise-supported |
Gemini and Gemini AI must not be used to process, store, or transmit Restricted Health Information (RHI), Protected Health Information (PHI), or any other sensitive data. | August 5, 2025 |
| Ubiqus / Acolad |
Approved up to SRDS moderate protection level. | Transcription/translation service | Paid | No SRDS high protection level level data use approval. | June 30, 2025 |
| PhoenixAI | https://genai.uchicago.edu/phoenixai-service-usage-guidelines | General Use | Enterprise-supported | Can be used for sensitive information with IRB approval. |
June 30, 2025 |
| REV | Approved up to SRDS moderate protection level. |
Transcription service that uses both human and AI (speech recognition). |
Paid | No SRDS high protection level level data use approval. |
June 30, 2025 |
| Perplexity.ai | Approved up to SRDS low protection level. | An AI-powered search and answer engine that combines large language models with live web search to deliver conversational responses. | Paid | Free version may log and use identifiable queries for model improvement. Paid version has a stronger privacy policy and does not log identifiable quieres for model improvement. | July 18, 2025 |
| Slack AI | Approved for data that is made publicly available by its source. |
AI in Slack can summarize channels and threads and serve up daily recaps. |
Paid | AI-generated outputs (e.g., summaries, recaps) may be posted in shared channels or viewed by multiple users. These outputs may surface, rephrase, or summarize others' messages, potentially highlighting content not intended for broader visibility. Slack’s privacy terms also permit use of data to improve AI features. Do not use with sensitive data. |
July 15, 2025 |
| AI tools embedded in Chrome extensions | Contact us to discuss your specific use case | Add-ons that provide AI-enhanced features such as writing assistance, summarization, or productivity enhancements directly within the browser. |
Free | Most Chrome extensions—AI or not—can capture browsing history, keystrokes, or sensitive data. Many are developed by unknown vendors and lack oversight or data security. Use of any extension should be carefully vetted. |
N/A |
| Otter AI | Not Approved - Inadequate Privacy protection levels. Use 3Play Media or Rev AI. | Transcription tool that uses AI to convert spoken language (from meetings, interviews, lectures) into written text. It also provides speaker identification, summarization, and collaboration features | N/A | Not approved from a security and privacy perspective. |
N/A |
| Orimon Chatbot | Not Approved - Inadequate Privacy protection levels. |
N/AAI-powered chatbot creation platform that enables users to build conversational agents that can handle customer interactions across multiple channels. | N/A | Not approved from a security and privacy perspective. |
N/A |
| Finalsite Chatbot | Not Approved - Inadequate Privacy protection levels. |
Automated tools embedded in websites to simulate conversation with users. They can answer questions, provide support, guide navigation, and even assist with transactions. |
N/A | At this time, the University has not approved the use of third-party chatbot products due to unresolved security and privacy concerns. Many of these tools do not align with our compliance requirements—including FERPA and GDPR—and may introduce vulnerabilities within our digital infrastructure. | N/A |
| Dante Chatbot | Not Approved - Inadequate Privacy protection levels. |
Automated tools embedded in websites to simulate conversation with users. They can answer questions, provide support, guide navigation, and even assist with transactions. |
N/A | At this time, the University has not approved the use of third-party chatbot products due to unresolved security and privacy concerns. Many of these tools do not align with our compliance requirements—including FERPA and GDPR—and may introduce vulnerabilities within our digital infrastructure. | N/A |
| You.com | Not Approved - Inadequate Privacy protection levels. |
AI-powered search engine and assistant that combines real-time web results with language model summaries. |
N/A | Privacy policy allows sharing data with third-party partners unless users log in and opt out. Tool collects query data for improvement by default. No SOC 2, no ISO 27001 (despite corporate SOC 2 Type 2 claim in blog). | N/A |
| Cleverbot | Not Approved - Inadequate Privacy protection levels |
A conversational AI chatbot designed for casual dialogue. |
N/A | Outdated architecture, no modern privacy protections, lacks University required security controls, and conversations are logged without user controls or disclosures. | N/A |
| Character.ai | Not Approved - Inadequate Privacy protection levels |
Platform that allows users to engage in open-ended conversations with fictional or user-created personalities | N/A |
Does not provide SOC 2, ISO 27001, or other enterprise attestations. Does not comply with GDPR/CCPA. Data is stored indefinitely. | N/A |
| Replika.ai | Not Approved - Inadequate Privacy protection levels |
A personal AI companion designed to simulate empathetic conversation. |
N/A | Designed for personal use and not aligned with University security standards; collects significant amounts of personal information and is not built with privacy-by-design principles. No SOC 2; though it received ISO 27001 back in 2017, no recent attestations; has had regulatory actions. Missing up-to-date GDPR/CCPA certification. |
N/A |
| Read.ai | Not Approved - Inadequate Privacy protection levels |
An AI that transforms meetings, emails, and messages into summaries. | N/A | Privacy and security concerns, including unauthorized recording of meetings, excessive access to calendars, and unclear data retention and deletion practices. No SOC 2, no ISO 27001, and no GDPR/CCPA compliance. |
N/A |
| Securly.ai | Not Approved - Inadequate Privacy protection levels | An AI tool used in K–12 environments to monitor student activity for signs of self-harm, bullying, or threats. |
N/A | Does not comply with University data retention standards; poor Bitsight Security score. No SOC 2, no ISO 27001, and no GDPR/CCPA compliance | N/A |
Generative AI Support
- Support
For support, submit a request to the IT Services Self-Service Portal or call 773.702.5800 for immediate assistance.
- Generative AI Tool Review
If you plan to acquire or use any generative AI technology tools, add-ons, or modules with University confidential data, please complete the Generative AI Tool Review form.