Research funders / U.S. National Science Foundation (NSF)
NSF AI policy for proposals
Yes, NSF lets you use generative AI to help prepare a proposal, and it encourages, but does not require, you to say in the project description how much AI you used and how. You stay responsible for the accuracy and authenticity of everything you submit. Since December 8, 2025, NSF's research misconduct definition explicitly covers fabrication, falsification or plagiarism done with AI-based tools. NSF reviewers must not upload proposal content to non-approved AI tools.
Independent guide, not an official U.S. National Science Foundation (NSF) page: we aren't affiliated with or endorsed by U.S. National Science Foundation (NSF). Summarised from its official guidance, last checked 21 November 2026. Rules can change, so your funding call wins if it says something different · See the full rules
Write yours
U.S. National Science Foundation (NSF) doesn't publish a template for this. If you want to note your AI use, this writes a short general statement you can adapt.
Worked example
Illustrative example: We used [AI tool] to edit the wording of parts of this project description and to suggest a shorter summary. We did not use it to generate research ideas, data or references. The PI reviewed all AI-edited text and takes responsibility for the content.
The rules at a glance
- When you need one
- Never strictly required. Proposers are encouraged to indicate in the project description the extent to which, if any, generative AI technology was used and how it was used to develop their proposal (NSF notice of December 14, 2023, page last updated May 21, 2026).
- Framework
- NSF Notice to Research Community: Use of Generative Artificial Intelligence Technology in the NSF Merit Review Process (December 14, 2023); PAPPG 24-1 Supplement 1 (NSF 26-200, effective December 8, 2025), Chapter XII.C research misconduct definition.
- Where it goes
- The project description.
- Must include
- whether generative AI was used, and to what extent (encouraged, not required)
- how it was used to develop the proposal (encouraged, not required)
- Also note
- Proposers are responsible for the accuracy and authenticity of their proposal, including content developed with AI help
- Research misconduct now explicitly includes fabrication, falsification or plagiarism done through AI-based tools, in proposing or performing research, reviewing proposals, or reporting results (PAPPG 24-1 Supplement 1, effective December 8, 2025)
- Reviewers must not upload any content from proposals, review information or related records to non-approved generative AI tools; NSF treats this as a breach of the confidentiality pledge
- Reviewers may share publicly available information with generative AI tools
- GRFP applicants must certify that their Personal, Relevant Background, and Future Goals Statement and Graduate Research Plan Statement are their own original work
Questions about U.S. National Science Foundation (NSF)'s rules
- Can I use ChatGPT to write my NSF proposal?
- NSF does not ban it. Its notice says proposers are responsible for the accuracy and authenticity of their submission, including content developed with generative AI tools. AI can create fabrication, falsification or plagiarism risks, and those count as research misconduct. Source
- Do I have to disclose AI use to NSF?
- It's encouraged, not required. NSF encourages proposers to say in the project description whether generative AI was used, to what extent, and how it was used to develop the proposal. Source
- Does AI misuse count as research misconduct at NSF?
- Yes. Since December 8, 2025, NSF's definition of research misconduct says fabrication, falsification or plagiarism counts whether you do it directly or through tools, including AI-based tools. That covers proposing or doing research, reviewing proposals, and reporting results. Source
- Can NSF reviewers use AI?
- Not on proposal material. NSF reviewers are prohibited from uploading any content from proposals, review information and related records to non-approved generative AI tools, and NSF treats it as a breach of the confidentiality pledge. They may share publicly available information with AI tools. Source
- Can I use AI on my NSF GRFP essays?
- NSF has published no GRFP-specific AI guidance that we could find. The GRFP solicitation does require you to certify that your Personal, Relevant Background, and Future Goals Statement and your Graduate Research Plan Statement are your own original work, and it reminds applicants that serious failures of scholarship and attribution can lead to research misconduct findings. Source
More help
Sources
- https://www.nsf.gov/policies/ai/merit-review
- https://www.nsf.gov/policies/document/pappg24-1-supplement-1
- https://www.nsf.gov/policies/pappg
- https://www.nsf.gov/funding/opportunities/grfp-nsf-graduate-research-fellowship-program/nsf26-526/solicitation
Your funding call wins if it says something different. Spotted a change? Tell us.