Introduction
Artificial Intelligence companies, notably OpenAI, have gained significant media attention thanks to their generative AI programs. For instance, ChatGPT and Perplexity provide text responses resembling human language, and DALL-E produces almost any kind of image – realistic, geometric, or abstract – based on written prompts. However, AI's scope extends beyond language and image-based chatbots. This libguide delves into a wide range of generative AI applications, with a particular emphasis on healthcare and research.
Secure Options to Work with Generative AI at NYU Langone Health
- UltraVioletAI – NYU Langone Health's private, secure, HIPAA-compliant ChatGPT-like tool. Access newer models, upload documents, organize chats, and more.
- Copilot.microsoft.com – Secure but NOT HIPAA-compliant. Use with proprietary information if logged in with Kerberos ID. Do NOT use with PHI.
Sharing Your Prompts
Please share your useful prompts on the NYU Langone Prompt Exchange, a community space for sharing prompts and discovering new ways to use AI systems.
Legal Compliance
Using generative AI applications must be fully compliant with policies and law:
- Do NOT use public AI tools for clinical documentation, including medical records or letters.
- Do NOT input PHI or other legally protected information into public AI applications.
- Do NOT use public AI tools with clinical or human subjects research data, even if de-identified.
- Do NOT disclose confidential business information to public AI tools.
- Do NOT allow public AI tools to record or upload internal NYU Langone meetings.
- Do NOT rely solely on AI output for work at NYU Langone Health; verify all results independently.
Additional Factors
- Access to NYULH’s private generative AI environment is rolled out gradually.
- Priority is given to innovation/research projects, mentored exploration, or exploration projects.
- Users receive email notifications upon approval and must sign a data use agreement.
Guidelines and Usage Policies
- Approved AI Tool: Apply for the NYU Langone GPT platform.
- Data Security: Obtain institutional permissions before adding sensitive data to AI programs.
- Avoid Other Tools: Do not use chatbots for NYU Langone Health work to ensure data security and ownership.
- Explore NYULH: Learn How NYU Langone Health uses generative AI.
For questions about AI at NYU Langone Health, please email: DAAIT@nyulangone.org
Quick Reference Guide
Download our convenient Quick Reference Guide for easy access to tips and resources.