Duke University provides a suite of AI platforms designed to make AI more accessible, secure, and practical for students, staff, and researchers. The suite will include ChatGPT, DukeGPT, MyGPT Builder and the AI Gateway, for advanced AI exploration. Each platform will provide hands-on tools and a structured, secure infrastructure for experimenting with AI models, developing AI-driven solutions, and integrating AI into Duke’s academic and research environments. Both new and advanced users will have access to flexible options and training.
DukeGPT
DukeGPT provides the university community with a secure, Duke-managed platform to explore and compare advanced AI models. Combining on-prem, open-source options with cloud-based foundation models, it ensures maximum privacy and robust data protection. Not to be used with PHI.

ChatGPT at Duke
ChatGPT can answer questions, write, edit, synthesize data and ideas, and even inspire. It provides conversational access to powerful language models like GPT-4o, and includes tools like a code interpreter, image generation, and web browsing. It is appropriate for use with some sensitive information, but not PHI.
MyGPT Builder
MyGPT Builder enables faculty, researchers, and staff to create and manage their own customized AI environments. By provisioning a dedicated OpenWebUI instance, individuals and groups- such as classes, departments, and administrative units- gain full administrative control to experiment within a secure, sandboxed environment.
AI Gateway FOR DEVELOPERS
The AI Gateway provides Duke developers and researchers access to LiteLLM API keys, enabling them to integrate AI capabilities into their own applications. It empowers users to build and deploy unique AI-driven tools, chatbots, and services, leveraging both on-prem and cloud-based models.
Microsoft Copilot
Microsoft Copilot
An AI companion, or chatbot, Microsoft Copilot can help boost productivity, spark creativity, and change the way you get things done. It is available at no cost to all members of the Duke community with a NetID. Copilot for Microsoft 365 is a paid service designed to enhance the way you work within the Microsoft 365 suite.
Which Tool Is Right For Me?
What are the benefits and differences between ChatGPT and DukeGPT? Here’s a quick guide: use ChatGPT for general, public tasks, and DukeGPT for anything tied to Duke or requiring extra privacy. Note neither tool is approved for sensitive data (e.g., PHI).
Here are the key benefits of each:
ChatGPT
ChatGPT is openAI’s commercial AI chatbot, available through Duke’s educational license.
Best for:
General-purpose tasks — writing, research, coding, brainstorming
creative or public-facing queries
Key features:
- Access to latest OpenAI models (e.g., GPT-4o)
- Built-in tools: code interpreter, image generation, file uploads
- Not linked to Duke systems
DukeGPT
This is Duke’s campus-specific AI tool, hosted and managed by Duke OIT for enhanced privacy and integration.
Best for:
- Academic or research support
- Duke-specific questions, internal tools, sensitive tasks
Key features:
- Data stays private within Duke’s systems
- Access to Duke-specific knowledge (e.g., directory, teaching tools)
- Multiple AI models, including open-source options
- Safe for institutional and educational use
Create a self-contained virtual resource with MyGPT Builder, running a chatbot to use for your own academic, work, or personal projects.
Available to: Students, Faculty, and Staff (university)
Pricing: Determined by needs

Copilot is powerful, built on Large Language Models (LLMs) such as GPT-4, or advanced tools designed to predict and generate text based on your internal data (emails, reports, presentations, articles, etc.).
Available to: Students, Faculty, and Staff (university and health system)
Pricing: Free
Access: Microsoft Copilot (login with NetID)
The AI Gateway for Developers allows you to build your own AI services using LiteLLM as the AI back-end (access the proxy service offered by LiteLLM [cloud or BYOkey]).
Self-Guided Training
This introductory course is designed for beginners with no prior knowledge of generative AI. You will start by gaining a high-level understanding of what generative AI is and how it works. Through interactive lessons and hands-on examples, you will learn fundamental skills like providing effective prompts and iteratively improving the generated outputs. As the course progresses, you will dive deeper into specific major generative AI models, including their unique capabilities and limitations. Finally, you will get practical experience using leading systems like GitHub Copilot, DALL-E, and OpenAI to generate code, images, and text. By the end, you will have developed core knowledge to start experimenting with generative AI in a responsible and effective way for a variety of applications. This course aims to provide a friendly introduction to prepare complete beginners for further exploration of this rapidly evolving technology.
This 15-minuute video introduces ChatGPT features including interface, custom settings, project management, model selection and lecture notes/citations.
https://academy.openai.com/public/videos/introduction-to-chatgpt-edu-2025-03-20
This course introduces students to the patterns and approaches for writing effective prompts for large language models. Anyone can take the course and the only required knowledge is basic computer usage skills, such as using a browser and accessing ChatGPT. Students will start with basic prompts and build towards writing sophisticated prompts to solve problems in any domain. By the end of the course, students will have strong prompt engineering skills and be capable of using large language models for a wide range of tasks in their job, business, personal life, and education, such as writing, summarization, game play, planning, simulation, and programming.
https://www.coursera.org/programs/duke-university-on-coursera-obsio/learn/prompt-engineering
The course will show how you can:
- Leverage prompt engineering techniques to generate more reliable outputs
- Master methods to verify and validate outputs
- Frame problems in alternative ways to reduce risk
- Apply generative AI for creative ideation
- Use Generative AI in ways that augment rather than replace human reasoning and creativity
https://www.coursera.org/programs/duke-university-on-coursera-obsio/learn/trustworthy-generative-ai
Learn how AI can break down complex topics, support structured thinking, deep research, essay outlines and other study aids.
Resume tips and interview preparation.
https://academy.openai.com/public/videos/ai-career-prep-resumes-and-interviews-2025-03-20
Throughout this series, learners will explore key topics including Explainable AI (XAI) concepts, interpretable machine learning, and advanced explainability techniques for large language models (LLMs) and generative computer vision models. Hands-on programming labs, using Python to implement local and global explainability techniques, and case studies offer practical learning. This series is ideal for professionals with a basic to intermediate understanding of machine learning concepts like supervised learning and neural networks.
https://www.coursera.org/specializations/explainable-artificial-intelligence-xai
This 17- minute video demonstrates how AI can automate and organize tasks and how to build a GPT.
What is the difference between each service?
Feature | ChatGPT | DukeGPT | MyGPT Builder |
AI Gateway |
---|---|---|---|---|
Target Audience | All Duke University users | All Duke University users | University faculty, researchers, staff |
Developers, advanced users |
Primary Use Case | General purpose chatbot use | Exploring AI models, learning | Custom AI environments, course-/department-specific tools |
Building AI-powered apps and tools |
Access Level | Standard ChatGPT interface | Limited feature set, predefined models | Full admin control, dedicated instance |
API key access, full customization |
Customization | None (standard ChatGPT) | Minimal (can use existing tools/pre-built GPTs) | High (build your own tools, knowledge base, etc.) |
High (API integration, model choice) |
Cost Structure | Pre-paid, or reduced cost | No cost for basic use | Cost-recover for some models, usage-based |
Cost-recovery service, usage-based |
Model Access | OpenAI's ChatGPT only (GPT-4o or equivalent base model) | Curated selection of open-source and cloud models | Curated selection of open-source and cloud models, others by request |
Curated selection of open-source and cloud models, others by request, BYOKey for cloud |
Security & Privacy | High (governed by institutional agreement) | High (Duke-managed, data-protected) | High (Dedicated, sandboxed) |
High (Secure API, user-managed) |
Video
An overview of DukeGPT.
Other Duke AI Resources
-
Duke Science & Technology - Exploring the benefits, ethics and potential impacts of AI on our community
-
Society-Centered AI Initiative - Designed to incubate research on the co-evolution of artificial intelligence and human behavior.
-
Let's Not Know Together - A collection of voices of students and educators who are collectively making sense of how AI fits—if at all—in their learning journeys.
For additional types of support and policy information, visit AI at Duke:

For service-specific inquiries, email aisuitesupport@duke.edu.