Approved AI Services
Card View | Matrix View
| Service | Description | Provider | Available to | Risk Category | Use Cases | Cost | Notes |
|---|---|---|---|---|---|---|---|
| AI Catalyst | AI Catalyst offers Stanford community members a structured pathway to develop customized AI solutions for research, teaching, administrative needs, and more. | UIT | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Custom Applications | Free | |
| AWS Bedrock | Provides access to Amazon AI and other third-party AI models. | Amazon | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Custom Applications, Developer and IaaS AI Tools | Paid | Approved for all risk levels when used with:
|
| AWS Q | A managed generative AI-powered chat assistant. | Amazon | Staff, Faculty, Students | Low Risk | Custom Applications, Developer and IaaS AI Tools | Paid | |
| AWS Sagemaker | A managed platform for building, training, and deploying machine learning models. | Amazon | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Custom Applications, Developer and IaaS AI Tools | Paid | Approved for all risk levels when used with:
|
| Azure OpenAI | Provides access to Azure OpenAI and other third-party AI models. | OpenAI | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Custom Applications, Developer and IaaS AI Tools | Paid | Approved for all risk levels when used with:
|
| Carina 2.0 | Secure computing cluster supporting both batch and interactive workloads. Today best suited for inference and modest training runs but upgrades are planned to add more model training capabilities. Local models supported. | Stanford Research Computing | Staff, Faculty | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Large-Scale Computing | Free, Paid | |
| ChatGPT Codex | An AI agent from OpenAI designed to perform real work across files, tools, and workflows — including gathering information, creating documents, and automating workflows, beyond basic coding assistance. | OpenAI | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI | Coding Assistance | Paid | Approved for use with Stanford data when logged in via Stanford SSO. Not approved for personal accounts or accounts without SSO. |
| ChatGPT on AI Playground | An AI chat assistant that can answer questions, write and edit content, analyze documents, and help with code. Explore ChatGPT's range of models, from lightweight options for everyday tasks to advanced models for complex work, in the Stanford AI Playground. | OpenAI | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with no PHI | Chat and Collaboration | Free | High-Risk non-PHI only; PHI review in progress for AI Playground. |
| ChatGPT through AI API Gateway | Designed for Stanford community members who want to develop their own AI solution(s) using the models available in the AI Playground. | OpenAI | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with PHI | Custom Applications | Paid | |
| Claude Code | A command line and development environment tool powered by Anthropic's Claude models that assists with code generation, refactoring, debugging, and repository-level editing. | Anthropic | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI | Coding Assistance | Free | Approved for use with Stanford data when logged in via Stanford SSO. Not approved for personal accounts or accounts without SSO. PHI processing is not covered under this approval. |
| Claude Cowork | An AI assistant that works directly on your computer, handling complex tasks like organizing files, creating documents, and executing workflows autonomously. | Anthropic | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI | Custom Applications | Free | Approved for use with Stanford data when logged in via Stanford SSO. Not approved for personal accounts or accounts without SSO. PHI processing is not covered under this approval. |
| Claude for Education | Stanford provided service for Anthropic's AI tool that can summarize, generate content, and assist with research or writing. Available as a browser-based tool, desktop application, and mobile app. Fable 5 remains disabled in the Stanford enterprise environment. Please see note on service page. |
Anthropic | Staff, Faculty, Students, Postdocs | Low Risk Moderate Risk High Risk with no PHI | Chat and Collaboration | Free, Paid | Approved for use with Stanford data when logged into your Stanford provided account via Single Sign On. This tool is powered by similar Anthropic models as those available in Stanford's AI Playground. |
| Claude on AI Playground | An AI assistant that can write and edit content, analyze documents, assist with coding, and handle complex reasoning tasks. Explore Claude's range of models from efficient options for quick structured tasks to advanced models for long running tasks and advanced coding workflows in the Stanford AI Playground. | Anthropic | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with no PHI | Chat and Collaboration | Free | High-Risk non-PHI only; PHI review in progress for AI Playground. |
| Claude through AI API Gateway | Designed for Stanford community members who want to develop their own AI solution(s) using the models available in the AI Playground. | Anthropic | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with PHI | Custom Applications | Paid | |
| Cline through the AI API Gateway | An AI-powered coding assistant that integrates with local development environments like VS Code, to generate, modify, and execute code. | Cline | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with no PHI | Coding Assistance | Paid | |
| Deepseek on AI Playground | An AI assistant that excels at coding, technical reasoning, advanced math, and complex problem-solving. Available in several models in the Stanford AI Playground, DeepSeek has an "Open Thinking" feature that lets you see the model's reasoning in real-time as part of its response. | Deepseek | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with no PHI | Chat and Collaboration | Free | High-Risk non-PHI only; PHI review in progress for AI Playground. |
| Deepseek through AI API Gateway | Designed for Stanford community members who want to develop their own AI solution(s) using the models available in the AI Playground. | Deepseek | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with PHI | Custom Applications | Paid | |
| Gemini on AI Playground | An AI assistant strong in reasoning, coding, multilingual tasks, and multimodal analysis. Explore Gemini models in the Stanford AI Playground, from lightweight options for quick responses to Gemini 2.5 Pro, a model for deep research and complex web development. | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with no PHI | Chat and Collaboration | Free | High-Risk non-PHI only; PHI review in progress for AI Playground. |
|
| Gemini through AI API Gateway | Designed for Stanford community members who want to develop their own AI solution(s) using the models available in the AI Playground. | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with PHI | Custom Applications | Paid | ||
| GitHub Copilot | An AI coding assistant that suggests code completions, functions, and documentation directly in the selected integrated development environment (IDE). | Microsoft | Staff, Faculty, Students | Low Risk Moderate Risk | Coding Assistance | Paid | Approved only for Cardinal Cloud GitHub. |
| Google Gemini Chat | Free, basic AI chatbot powered by Google Gemini AI models. It can answer questions, write content, summarize documents, and generate images, but it is not integrated into Google Workspace. | Staff, Faculty, Students | Low Risk Moderate Risk | Chat and Collaboration | Free | ||
| Google Gemini Enterprise Pro | Upgraded tool for users with heavier research, productivity, or workflow needs. Provides higher usage limits and access to more powerful reasoning models for vast datasets at a discounted rate. | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Chat and Collaboration | Paid | ||
| Google Gemini Enterprise Standard | Gives you access to Google's latest AI models within the productivity tools you already use. | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Chat and Collaboration | Free | ||
| Google NotebookLM | AI-powered research and note-taking tool that lets you upload documents and ask questions about them. Uses Google Gemini models to summarize content, extract insights, and generate new material. | Staff, Faculty, Students | Low Risk Moderate Risk | Chat and Collaboration | Free | ||
| Google Vertex | Provides access to Google AI and other third-party AI models. | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Custom Applications, Developer and IaaS AI Tools | Paid | Approved for all risk levels when used with:
|
|
| Grammarly | An AI writing assistant that integrates with browsers, Microsoft Office, Google Workspace, and other applications to help with grammar checking, style improvement, tone suggestions, and AI-generated text. | Grammarly | Staff, Faculty, Students | Low Risk Moderate Risk | Chat and Collaboration | Paid | |
| Llama-4 on AI Playground | An open-source AI assistant strong in writing, summarization, translation, and multilingual tasks. Explore Llama models in the Stanford AI Playground, from fast, lightweight options to more capable versions for creative tasks and research support across languages. | Meta | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with no PHI | Chat and Collaboration | Free | High-Risk non-PHI only; PHI review in progress for AI Playground. |
| Llama-4 through AI API Gateway | Designed for Stanford community members who want to develop their own AI solution(s) using the models available in the AI Playground. | Meta | Staff, Faculty, Students, Affiliates | Low Risk Moderate Risk High Risk with PHI | Custom Applications | Paid | |
| Marlowe | A large-scale, high-performance, GPU-based computational instrument built on NVIDIA DGX H100 Superpod architecture and designed to deliver cutting-edge computational performance. Optimized for large scale model training. | HAI, Stanford Research Computing | Staff, Faculty | Low Risk Moderate Risk | Large-Scale Computing | Free, Paid | |
| Microsoft 365 Copilot | A more powerful AI assistant (powered by OpenAI models) built directly into Microsoft 365 applications like Word, Excel, Outlook, PowerPoint, and Teams for increased functionality. | Microsoft | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Chat and Collaboration, Transcription and Meeting Based | Paid | |
| Microsoft Copilot Chat | A more advanced AI chatbot that runs on Microsoft’s infrastructure. Provides chat style interactions and can generate text and images but is not integrated into Microsoft 365 applications (Outlook, Word, Excel, and PowerPoint). | Microsoft | Staff, Faculty, Students | Low Risk Moderate Risk | Chat and Collaboration | Free | |
| NotebookLM Enterprise | Enterprise version of Google's AI-powered research and note-taking tool, enabling document upload and Q&A with added enterprise-grade data governance. | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Chat and Collaboration | Paid | Distinct from the standard "Google NotebookLM" entry, which remains approved only for Low/Moderate Risk. |
|
| OpenAI ChatGPT Edu | OpenAI's AI tools that can generate text, answer questions, create images, write code, and assist with a wide variety of tasks. Available as browser based tool, desktop application, and and mobile app. |
OpenAI | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Chat and Collaboration | Free, Paid | Approved for use with Stanford data when logged into your Stanford provided account via Single Sign On. This tool is powered by similar OpenAI models as those available in Stanford's AI Playground. |
| SHC-SoM Secure GPT | Stanford Medicine hosted platform that allows users to test out multiple AI models in a safe, Stanford Medicine supported environment. | OpenAI | School of Medicine | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Chat and Collaboration | Free | |
| Sherlock | A shared computing cluster providing access to a base set of managed computing resources, GPU-based servers, and a multi-petabyte, high-performance parallel filesystem for short-term storage. Suitable for model training. Numerous local models supported. | Stanford Research Computing | Staff, Faculty | Low Risk Moderate Risk | Large-Scale Computing | Free, Paid | |
| Slack AI Tools | Brings built-in AI features to Stanford's instance of Slack to summarize threads, surface answers, automate routine workflows, and more. | Slack | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI | Chat and Collaboration | Free | |
| Voice commands in Google Docs and Slides | A built-in Google Docs and Slides feature that lets you type and edit documents, speaker notes, and captions using your voice. Compatible with the latest versions of Chrome, Edge, and Safari. | Staff, Faculty, Students | Low Risk Moderate Risk High Risk with no PHI High Risk with PHI | Transcription and Meeting Based | Free | Approved for use with Stanford data when logged into your Stanford-issued Google account. |
|
| Zoom AI Companion | A native Zoom feature that boosts meeting productivity with AI-powered summaries, smart recordings, and the ability to answer questions during meetings. | Zoom | Staff, Faculty, Students | Low Risk Moderate Risk | Transcription and Meeting Based | Free |
