AI Usage in Proposals
As generative Artificial Intelligence (AI) tools have become more developed and increasingly commonplace, federal sponsors have implemented rules and regulations surrounding the use of AI in research proposals, including mandatory disclosure requirements in certain cases.
What is Generative AI?
Generative AI is a type of artificial intelligence that can create new content, such as text, images, audio, video, computer code, or data, based on patterns it learned from large amounts of existing information. As defined by the Department of Energy, generative AI “refers to AI models that emulate the structure and characteristics of input data in order to generate derived synthetic content, also referred to as “output,” including text, code, images, videos, audio, and other digital media” (https://www.energy.gov/documents/doe-p-2031-use-generative-ai).
Examples of generative AI include (but are not limited to):
- Text generation: ChatGPT, Microsoft Copilot, Claude
- Image generation: DALL·E, Midjourney
- Code generation: GitHub Copilot
- Audio/music generation: AI tools that create speech or music from prompts
Guidelines by Agency
Federal guidance on AI use in federally funded research is an evolving topic. The following provides current information from select agencies. The Office of Contracts and Grants (OCG) advises reviewing funding opportunity announcements and award terms and conditions for additional restrictions on AI usage.
- NSF permits the use of generative AI tools to assist in proposal preparation.
- Proposers remain fully responsible for the accuracy, originality, and integrity of their submissions.
- NSF encourages investigators to disclose how generative AI was used in developing the proposal.
- AI-generated content must comply with all research integrity requirements (see NSF Proposal & Award Policies & Procedures Guide (PAPPG)), and confidential proposal information should not be shared with public AI systems.
- For additional details, see: Notice to Research Community: Use of Generative Artificial Intelligence Technology in the NSF Merit Review Process.
- Effective September 25, 2025, “Applications that are either substantially developed by AI or containing sections substantially developed by AI are not considered the original ideas of applicants and will not be considered by NIH.”
- AI tools may be appropriate to assist in limited aspects of application preparation.
- NIH uses methods to identify AI-generated content and may take compliance or enforcement actions if inappropriate AI use is identified.
- For additional details, see: Apply Responsibly: Policy on AI Use in NIH Research Applications and Limiting Submissions per PI and Supporting Fairness and Originality in NIH Research Applications.
- “NEH will permit applicants to use AI technology to prepare their applications; however, NEH will require all applicants to specifically acknowledge (by use of footnotes or other marginal notations) when they have inserted AI-generated text into their proposals.”
- NEH reserves the right to declare an application ineligible if it does not comply with the disclosure requirement.
- Specifically, AI may be used as a writing aid, such as for drafting, editing, or language assistance.
- However, using AI does not relieve applicants of responsibility for the proposal's accuracy, clarity, or sourcing.
- For further details, see: Policy on the Use of Artificial Intelligence for NEH Grant Proposals.
- DOE currently does not have a department-wide grant proposal policy that bans or requires disclosure of generative AI use.
- However, DOE guidance permits the use of AI tools as an aid to writing and idea development while emphasizing that investigators remain fully responsible for the accuracy, originality, and integrity of proposal content.
- Applicants should carefully verify AI-generated information and avoid entering confidential or proprietary information into public AI systems.
- See also: DOE Use of Generative AI.
- “Currently, Science Mission Directorate (SMD) does not prohibit the use of Generative Artificial Intelligence (AI) tools, such as ChatGPT, or professional human writers in the preparation of proposals or required award reports” (https://science.nasa.gov/researchers/sara/faqs/#faq-32).
- In terms of citation, proposers and grantees must acknowledge when content is not the creative product of the proposal team.
- Any material contained in proposals (or in reports) to NASA that is not the product of the team must be cited.
- If AI is used, the name of the program, version number, the date and time, and a statement on how the Generative AI was used (https://science.nasa.gov/researchers/sara/faqs/).
CU Boulder Policies Governing AI Usage
- Use AI tools available and recommended for use on the CU Boulder campus found on the Office of Vice Chancellor for IT and Chief Information Officer website. If you have any questions about access to these tools, please email oithelp@colorado.edu.
- AI tool use is only approved for data classified as Public (see CU's Data Classification Standard for more information).
- Information shared with Generative AI tools using default settings is not private and could expose proprietary or sensitive information to unauthorized parties.
- Research proposals, plans and results are classified as “confidential” or, if they contain export-controlled information, “highly confidential.”
- Faculty, staff, and students using these tools must read and understand the AI tool vendor’s policies and become familiar with how the vendor plans to store, process, or handle CU’s data.
- See additional resources.