What funders say about AI in grant applications

AWHONN prohibits AI tools, including text generators, paraphrasers, summarizers, and translators, in preparing any scholarship, grant, or award application. AWHONN may check for AI use through human review and by requiring a signed declaration.

The Educating Character Initiative (ECI), the Program for Leadership and Character's grantmaking arm, does not approve the use of generative or assistive AI to write grant proposals. This appears in both its 2027 institutional grants RFP and its Communities of Impact grants RFP. We found no published guidance for reviewers.

Activate Fellowship applicants may use AI tools to help draft, edit, or brainstorm, but all concepts must be their own and they must be able to speak fluently about everything they submit. Technical reviewers may not use AI to write reviews or paste applicant materials into public AI tools, and Activate uses enterprise AI for research and summarization while humans make all selection decisions.

APF lets applicants use AI tools to organize their thoughts and structure a proposal, but the final ideas and content should be their own. Proposals developed primarily by AI will be declined. Applicants are asked to acknowledge and disclose AI use in the application.

APTF scholarship applicants are expected to submit original responses in their own words. The Board may disregard submissions that appear to be fully generated by AI tools.

ECRAC's FY2026-2027 Individual Artist and Culture Bearer grant guidelines say AI cannot be used to fully generate art, application answers, or art grant proposals. Applications with excessive AI-generated content may be deemed ineligible.

eRA is NIH's grants administration system, so NIH's policies apply. NIH will not consider applications substantially developed by AI, or containing sections substantially developed by AI, to be original ideas of applicants. NIH peer reviewers may not use generative AI to analyze applications or write critiques.

The Foundation for Physical Therapy Research's 2026 Research Grant guidelines say that, consistent with NIH, it does not consider applications substantially developed using AI to be original work, and applications with substantive and undue AI use will be disqualified from review. Its 2026 Postdoctoral Fellowship FAQ states that AI cannot be used to create or refine research or training plans.

The International OCD Foundation says generative AI should not be used to conceptualize a Research Grant Program proposal, though it may help with drafting, such as translating text when English is not the PI's first language. The PI must state in the application whether any AI tools were used and what that use involved.

For the Internet Society Fellowship Program, generative AI is not prohibited but should only be a supportive resource, such as for brainstorming, structural editing, translation, grammar correction, or accessibility. Essays and project proposals must reflect the applicant's own insights and experience, and plagiarism or uncredited AI-generated narratives lead to disqualification.

The Jerome Foundation says AI may help with tasks like proofreading, translation, or summarizing, but the substance, ideas, and concepts in an application must be the applicant's own. Applicants who submit AI-generated or AI-augmented work samples without disclosure will be deemed ineligible. Panelists are told not to enter applications into AI tools, and staff will not enter application content into AI tools except to transcribe video responses for accessibility.

Guggenheim Fellowship applicants must attest that any generative AI use in preparing the application was limited to help with brainstorming, researching, editing, or formatting, and that the substantive ideas, analysis, and conclusions are their own.

Mid-America Arts Alliance strongly discourages using AI to generate answers to narrative application questions, but allows AI to proofread and edit responses to fit word or character limits. Its Catalyze guidance does not allow AI-generated artwork, and the Artistic Innovations guidelines say applications using AI for other purposes will be deemed ineligible. Reviewers are prohibited from using AI tools.

The Mississippi Arts Commission's AI statement says a human applicant must develop, review, and enter all required data in a grant application, and a human must be the primary author of any artwork. Undisclosed use of generative AI makes an application ineligible.

NIH will not consider applications, or sections of applications, that are substantially developed by AI to be the original ideas of applicants. AI may help with limited aspects of preparing an application, but you are responsible for plagiarism, fabricated citations, or other misconduct it introduces. NIH peer reviewers may not use generative AI to analyze applications or write critiques.

Nevada Arts Council applicants may use AI in writing applications or creating work samples, but these may not be fully AI generated and must have meaningful human authorship. Applicants must disclose if and how they used AI. Staff and panelists may not use AI to review applications or generate feedback.

PERF acknowledges that some investigators use generative AI when writing proposals, but expects your letter of intent (LOI) to reflect your own work. LOIs that do not follow this will be returned without review.

For Phi Theta Kappa scholarships, ChatGPT is a tool and is not meant to be the author of your essay. PTK treats copy-pasted AI-produced work as plagiarism, which removes an application from consideration, and asks applicants to verify their essays are their own work.

For SCA Research Grants, any use of AI tools to prepare the application must be disclosed and described on the title page, and AI tools may not be used to generate substantial segments of content. Concerns that generative AI was misused will be investigated. Grant reviewers may not use AI tools that require entering grant contents or other confidential review material.

The South Carolina Arts Commission's AI policy allows AI for editing, proofreading, or improving the clarity and quality of narrative text, but applicants who use AI to help write or edit their application do so at their own risk. Generative AI must not be used to create the proposed artwork or to generate work samples.

Spencer lets applicants use generative AI as an assistive tool, but not to submit verbatim AI-generated drafts, and any generative AI use must be disclosed through a checkbox and short summary in the online submission system. Using AI only for grammar, spelling and punctuation does not need disclosure. Reviewers may not use generative AI to summarize, analyze or otherwise help evaluate proposals.

UKRI restricts how applicants use generative AI: you must not use it to generate an entire application, or sections of one, without human involvement, and must not use it in interviews. You are expected to be transparent where you have used it (for example idea generation, data analysis, comparing literature, code or an abstract), and UKRI says this will not affect assessment. Minimal use such as translation or improving your English does not need to be disclosed.

Whitley Fund for Nature wants original work: all concepts, ideas, and activities in the application must come from the applicant, their team, or stakeholders, and applicants are asked not to submit content heavily reliant on AI. Acceptable uses include clarifying terminology, translating to English without changing content, making concept graphics (not photos), and correcting language or grammar. Applicants who used generative AI should detail which software, for what purpose, and how AI content was adapted to their project.

The Wilson Ornithological Society does not allow generative AI to generate text for its research grant or travel/caregiver award applications. Generative AI may be used to translate text the authors wrote into English, but this must be disclosed. Committees and reviewers may not enter any application material into generative AI tools.

For DivCHED Strategic Projects Funding (SPF) proposals, applicants may use assistive, non-generative AI (grammar, spelling, formatting, readability) without disclosure, but must disclose any generative AI used to prepare the proposal or planned for the proposed work. Generative AI is prohibited in proposal review.

ACS allows AI tools in grant applications, but you are responsible for all content, including anything produced by an AI tool or large language model. ACS says you should credit any AI tool used to develop your application and cite sources where possible. You should follow your institution's AI rules, and your use cannot conflict with the ACS Guidelines for Research and Peer Review Integrity.

AHA lets applicants use a large language model such as ChatGPT, or another AI tool, to create or edit proposal content, but you must disclose this when you submit. AHA says disclosure does not affect peer review, and a proposal may be administratively withdrawn if AI use is not disclosed accurately. Reviewers may not use AI to write or edit critiques, or upload any part of a proposal into an AI tool.

AIA requires award applicants to disclose any use of AI in creating their submissions. The application portal asks whether AI was used, which tools, and which parts of the application were affected.

The American Lung Association requires you to briefly describe any AI tools you used and why as part of your submission. You may use AI-based tools for limited administrative help, such as grammar editing, formatting, or readability, but AI must not be used to generate falsified data, citations, images, or research results. Failing to disclose material AI use can lead to withdrawal of the application, rescission of an award recommendation, or termination of funding.

AOTF does not prohibit AI tools for help writing proposals (such as spell-check and language suggestions), but applicants are fully responsible for accuracy and citations and must acknowledge any AI-generated text in footnotes. Reviewers may not upload AOTF applications into any AI system.

FWF generally permits AI in preparing proposals, but significant AI contributions to design or drafting must be disclosed, and the PI Projects guidelines ask for a brief description of any AI use in the Co-authors form. Applicants are fully responsible for all content, and FWF checks applications and reviews for AI use at random and on suspicion.

The Biochemical Society allows applicants to use AI to help write or edit applications and to collect and analyze data, but they must disclose which tool was used and how. AI may not be used to generate images or videos, and reviewers may not use AI to generate or summarize content or upload applications to AI tools.

Cancer Grand Challenges' Expression of Interest guidelines require applicant teams to correctly and explicitly attribute outputs from generative AI tools in their applications. Teams should list the AI source, name the specific model and software where practicable, and say how the content was generated (for example, the prompt used).

Cancer Research UK advises caution with generative AI, and applicants who use it must explicitly attribute AI outputs in the application, where practicable naming the model and software and saying how the content was generated. Applications must contain accurate information, and host institutions must confirm the application does not breach CRUK research integrity rules. Peer reviewers must not share application content with generative AI tools or use them to formulate or edit their critiques.

The Fibrolamellar Cancer Foundation permits grant applicants to use AI tools in developing proposals, but all AI use and how it was used must be disclosed in the application. Applicants are fully responsible for accuracy. Review committee members may not use AI tools to analyze, summarize, or interpret applications or to write critiques.

The Firebird Foundation expects grant applications to be the applicant's own work but encourages responsible use of AI. Applicants must disclose whether they used AI, which tools they used, and how.

FCA does not require disclosure when AI was used to help write text answers, though it notes AI has a flattening effect on artist statements and project narratives. Disclosure is required when AI generates media or is a primary contributor to the final work, by writing "Generated by AI" in each relevant work sample description. FCA staff and panelists do not use AI to review applications.

The Gerda Henkel Foundation generally permits generative AI when preparing applications, as long as applicants take full academic responsibility and follow good scientific practice. Applicants must document in the application form which stages of their work involved generative AI. Preparing an application entirely with AI, without the applicant's own academic review and control, is not permitted.

The DFG generally permits AI when preparing a funding proposal, and AI use does not affect how a proposal is assessed. Applicants must disclose which generative models were used, for what purpose, and to what extent. Reviewers may use AI only in a supporting role, with all substantive judgements resting with the reviewer and confidential proposal content protected.

Golden Field Office notices that use DOE's NOFO Part 2 version 4.0 (for example DE-FOA-0003548, Unleashing Tribal Energy Development) require you to state in the project summary whether and how you used generative AI to develop your application, and say you remain responsible for the accuracy, authenticity, and authorship of everything you submit. Notices using Part 2 version 2.1 (for example the 2026 geothermal field tests NOFO) instead require any AI use to be appropriately attributed. Check your NOFO's Part 2.

Grayson-Jockey Club Research Foundation requires full disclosure of any AI-assisted technology (such as ChatGPT or Claude) used in writing, editing, or creating images for an application. Assistive AI for grammar and style is allowed, but generating content with AI is strongly discouraged. Applicants must verify all AI-generated references, and fabricated references lead to immediate rejection.

HEI requires applicants to disclose AI use at the proposal development stage, including which tools were used, which tasks were AI-assisted, and how and to what extent. Proposals must include a statement declaring generative AI use. Reviewers may not upload proposals into AI tools or use AI to analyze, summarize, or evaluate proposal materials.

Heart & Stroke requires applicants to clearly state if and where material in their proposals or application materials was generated by AI, and to reference that material appropriately. Non-generative tools like Grammarly that only correct appearance, clarity, or presentation do not need to be disclosed. Heart & Stroke may use AI-based tools to help prepare Scientific Officer notes summarizing review committee discussions.

The Research Grants Council requires applicants to disclose, where applicable, their use of AI, including the tools used, the purpose, and how and to what extent AI was used. This appears in both the GRF/ECS call letter and the Theme-based Research Scheme call. TRS applicants must also follow their university's AI policies and take full responsibility for AI-generated content.

Indy Arts Council asks grant applicants to disclose if and how their submission used AI, and says submissions that do not meet its expectations will not be competitive. Reviewers may not use large language models or generative AI to analyze or write evaluations, and the council says it does not use AI anywhere in its grant processes.

IAFNS allows applicants responding to its requests for proposals to use generative AI, but requires disclosure. The principal investigator is responsible for the content of any submission. Its brain-health nutrition RFP also requires institutions to make sure AI outputs are valid and free of false or hallucinated information.

The Jack Kent Cooke Foundation expects scholarship application responses to be wholly the applicant's own and written from their own perspective. It recognizes generative AI could be useful for some parts of the application and asks applicants to report any AI use in a space provided in the application.

Kidney Cancer Association award guidelines permit language models and AI writing tools as a supplementary aid in preparing grant proposals, but core ideas, data, and content must be the applicant's original work. AI use should be transparently declared in the application, and applicants are responsible for verifying all AI-generated content.

The Lipedema Foundation's applicant FAQ says AI may be used to create a proposal with transparent disclosure, and points to the AI section of its RFP page for details.

Lloyd's Register Foundation allows responsible use of generative AI in grant applications, but expects it to be accurate, transparent, ethical, and lawful. Applicants must tell the foundation if they use generative AI to create or support any original content for an application, and they are responsible for how they use it.

The Massage Therapy Foundation permits some AI use, such as brainstorming, summarizing, or creating figures, but it must be fully disclosed, including the tool, purpose, and extent. Presenting AI-generated content as your own is prohibited, and primarily AI-generated submissions will not be reviewed. MTF does not enter submitted proposals into AI tools.

The Misophonia Research Fund (MRF) permits large language models to assist with proposal writing, but says critical thinking, review, and analysis must be done without AI. Applicants must disclose and describe any AI use. Disclosure will not hurt the review, but a proposal found to use AI without disclosure is subject to withdrawal.

The Missouri Arts Council strongly cautions against using AI to write grant applications, and any use of AI tools must be disclosed. Using AI for administrative tasks such as formatting, checklists, or grammar checks is generally permitted with no disclosure. Panel reviewers may not use AI to review applications.

NASA's Science Mission Directorate (SMD) does not prohibit generative AI in ROSES proposals, but any AI use must be cited in the proposal's references section with the tool name, version, date and time, and how it was used. Proposers are responsible for the accuracy of AI-assisted content. No published NASA rule on AI use by reviewers was found.

NASA's Science Mission Directorate (SMD) does not prohibit generative AI in ROSES proposals, but any AI use must be cited in the proposal's references section with the tool name, version, date and time, and how it was used. Proposers are responsible for the accuracy of AI-assisted content. No published NASA rule on AI use by reviewers was found.

NEH lets applicants use AI to prepare applications, but you must acknowledge any AI-generated text you insert, using footnotes or other marginal notes. NEH can declare an application ineligible if you do not. Peer reviewers may not upload NEH applications to outside AI or cloud services.

If you use generative AI in preparing a NIAS fellowship application, you must disclose it in the application and list it in your references, as set out in the regulations.

DOE's Office of Science applies a "fair-use" approach to generative AI. You may use AI tools to help present your own intellectual work, but you must disclose any AI use in the application unless the tool was used only to edit your own original draft. A project narrative that is not your own work, including a machine-generated one, may count as research misconduct.

Ohio Arts Council applicants may use AI to help write or edit grant proposals, but remain responsible for making sure everything is their own work and accurate. Applicants are asked to disclose any third-party AI used in an application or its attachments. OAC's policy also bars AI content from being the sole or determinative basis for eligibility, scoring, or award recommendations.

Applicants to Prevent Cancer Foundation research grants and fellowships must disclose any use of generative AI in preparing the application. Applicants are fully responsible for the content, and the disclosure is for information gathering only and does not affect review.

Applicants to the Royal Geographical Society's grants programme must disclose any use of generative AI tools, except for translation or improving English. Disclosure does not affect the assessment. Reviewers should not use generative AI tools when reviewing.

SOT's general AI guidance says that when AI is used to produce content shared with SOT leadership, volunteers, or others in the SOT community, its use must be documented with an attestation statement. SOT does not support content generated only by AI without human review, and does not recommend AI for content meant to be unique to an individual, such as a personal statement in an awards application. Acceptable uses may differ by program.

SDDOT research proposals must include an AI Use Statement describing how AI tools were used to prepare the proposal and how AI is expected to be used during the research. If no AI tools were used, the proposal must say so in a set sentence.

Applicants may use generative AI tools to prepare a Wellcome grant application but must report any substantive use on the application form. Minimal use, such as translation, improving English, or formatting, does not need to be disclosed. Expert reviewers and committee members must not enter confidential applications into generative AI tools.

AHRQ has no AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. NIH's AI policies are not stated to apply to AHRQ. Check each NOFO for added rules.

A*STAR allows applicants to use generative AI responsibly to prepare grant applications, as long as research integrity, accountability, and originality are preserved. Applicants must declare they are fully responsible for the content; this is not a mandatory AI disclosure unless the grantor requires one. Compliance checks may include AI detection tools.

AFOSR's open research BAA says generative AI use in proposals is not prohibited, and AFOSR strongly encourages (but does not require) you to disclose it. Any use is at your own risk, and you are responsible for avoiding plagiarism, misuse of intellectual property, and misinformation. External reviewers are prohibited from using generative AI.

Alabama Humanities Alliance grant guidelines acknowledge that applicants may write proposals with AI programs like ChatGPT or Claude. The guidelines ask those applicants to personally review the application, check the facts, and make sure AI has not produced dense, repetitive text or left out key details.

The Humboldt Foundation welcomes AI as an aid in the application process and does not currently require applicants to identify its use. In the Henriette Herz Scouting Programme, peer reviewers are asked not to use any generative AI tools when preparing their statements.

The Alzheimer's Association allows applicants to use large language models and other generative AI tools to prepare letters of intent and full applications. You are fully responsible for everything in your proposal, including parts produced by an AI tool. The Association says using these tools will not affect the review of your application.

AAUW lets fellowship and grant applicants use generative AI to help prepare their application narratives. AI should assist, not replace, your own ideas and writing, and narratives should reflect your authentic voice and original work. No disclosure requirement is stated.

ABTA allows generative AI in grant applications, with applicants fully responsible for the content and encouraged to review anything AI produces. For Metastatic Brain Tumor Collaborative grants, publishable content like the lay summary must be authored by the applicant or project team, and reviewers may not use AI to analyze, summarize, or critique applications.

For the ACS Petroleum Research Fund (PRF), disclosing AI use is optional, but if you disclose that you used AI tools to prepare your application, you must describe how you used them. PRF reviewers may not upload any proposal content to generative AI tools.

ACLS allows applicants to use AI tools to research or write a proposal, and the applicant stays responsible for any errors, hallucinations, or plagiarism. ACLS encourages, but does not require, applicants to say where and how they used AI. ACLS peer reviewers may not use generative AI to read applications or write reviews.

AWS does not prohibit using AI to help draft personal essay statements for its national scholarships. Reviewers look for authenticity and originality, so applicants are encouraged to share real experiences.

Atlas Network's grant application guide advises applicants not to rely on AI to come up with their goals, but to use it to pressure-test goals they have already drafted. It also advises using an AI account that does not train on or retain inputs before pasting in sensitive project details.

Bartlesville Community Foundation says AI can be a helpful tool for writing grant applications, but it should complement the applicant's own expertise and effort rather than replace them.

BHF says applicants can use generative AI when preparing an application, as long as the tools are used responsibly and in line with relevant legal and ethical standards. Committee members and external reviewers must not use generative AI when assessing grant applications.

CIHR follows the joint guidance it issued with NSERC, SSHRC and CFI. Applicants may use generative AI, including to draft, summarise or translate parts of an application, and there is no requirement to disclose it. The named applicant is accountable for everything in the application, and reviewers are strictly prohibited from using any generative AI tool, including institutionally procured ones.

The Community Foundation for Monterey County does not prohibit the use of AI in its application-based grant requests. If applicants use AI tools, CFMC encourages them to proofread before submitting, keep their organization's voice, and include local data and information to support their statements.

The Community Foundation of Henderson County welcomes applicants' use of AI tools for grant and scholarship applications, as long as applicants remain fully responsible for the originality, accuracy, and integrity of what they submit. Applications may ask whether and where AI was used, but the answer carries no weight in the committee's decision. Volunteer reviewers do not use AI to evaluate applications.

The Department of Education says using AI to develop grant application materials is allowable. You must still certify that everything in your application is true, complete, and accurate, whether or not AI generated it. ED has not published guidance on reviewers using AI.

The Department of Education says using AI to develop grant application materials is allowable. You must still certify that everything in your application is true, complete, and accurate, whether or not AI generated it. ED has not published guidance on reviewers using AI.

NSF allows proposers to use generative AI. It encourages, but does not require, proposers to say in the project description whether and how they used it, and proposers stay responsible for the accuracy and integrity of everything they submit. NSF reviewers may not upload any proposal or review content to non-approved generative AI tools.

The ERC Scientific Council has said that using AI to draw up a research proposal does not relieve the author of full and sole authorship responsibility, including for acknowledgements, plagiarism, and good scientific and professional conduct. Reviewers may not upload proposals to online AI tools or use AI to summarise or assess a proposal, but may use AI to polish language or search for information on the topic.

FDA has no AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. Check each FDA NOFO for added rules.

Humanities Nebraska says applicants can use generative AI to write grant applications. As with any application, responses should accurately reflect what the applicant can realistically accomplish.

Medical Research UK says applicants may use AI tools in developing their application where appropriate, and remain responsible for its accuracy, originality, and integrity. Generative AI should be used responsibly, and as best practice its outputs in applications should be acknowledged.

Mitacs allows applicants to use generative AI to develop, write, or edit proposals. Applicants who do are expected to follow the FASTER principles, avoid entering confidential or proprietary information into AI systems, and remain responsible for the accuracy and originality of their proposal. Reviewers may not put proposal content into generative AI systems.

The National Lottery Community Fund says you can use AI tools to help write a funding application, and it will not reject an application just because AI was used. Make sure the answers are accurate and reflect your own work.

The Novo Nordisk Foundation's Policy on Ethical Use of AI applies to grant applicants as well as reviewers. Applicants are not barred from using AI, but they remain fully accountable for anything created with AI assistance. Committee members and peer reviewers may not use AI to evaluate applications or upload any application content to an AI tool.

The NSF X-Labs solicitation and FAQ say nothing about AI use by proposers or reviewers. NSF's agency-wide notice on generative AI in merit review is the only published guidance: NSF encourages (but does not require) proposers to describe any generative AI use, holds proposers responsible for accuracy, and bars reviewers from uploading proposal content to non-approved AI tools.

The Department of Education says using AI to develop grant application materials is allowable. You must still certify that everything in your application is true, complete, and accurate, whether or not AI generated it. ED has not published guidance on reviewers using AI.

OESE has no AI guidance of its own. Its competitions, such as EIR Mid-Phase Grants, require applicants to follow the Department of Education's Common Instructions, which allow AI in developing application materials. You must still certify that the whole application is true, complete, and accurate, whether or not AI generated it.

The Department of Education says using AI to develop grant application materials is allowable. You must still certify that everything in your application is true, complete, and accurate, whether or not AI generated it. ED has not published guidance on reviewers using AI.

PCORI does not bar applicants from using AI tools to help prepare letters of intent or full applications, but applicants who do so act at their own risk. By submitting, applicants represent that the materials are their team's original ideas and content, and AI-introduced fabrication, falsification or plagiarism can be handled under PCORI's Research Misconduct Policy. PCORI prohibits putting confidential application information into tools such as generative AI during its review process.

The Research Council of Finland allows generative AI and similar tools to support application preparation and does not require you to say whether you used them. Staff, reviewers, and decision-makers may not enter application or review content into generative AI tools.

SSHRC follows the joint guidance it issued with CIHR, NSERC and CFI. Applicants may use generative AI, including to draft, summarise or translate parts of an application, and there is no requirement to disclose it. The named applicant is accountable for everything in the application, and reviewers are strictly prohibited from using any generative AI tool, including institutionally procured ones.

Suffolk Community Foundation lets applicants use AI tools such as ChatGPT or Copilot to help write funding applications and reports, and will not reject an application solely because AI was used. Applicants remain responsible for accuracy. Disclosure is not required, but the foundation expects openness where AI materially affects the content or delivery of a proposal.

The Tinker Foundation is open to applicants using AI tools to refine Institutional Grants applications, including for editing and translation. It discourages relying on AI to develop or describe project concepts, saying applications are strongest when they reflect the organization's own experience and understanding.

NSF allows proposers to use generative AI. It encourages, but does not require, proposers to say in the project description whether and how they used it, and proposers stay responsible for the accuracy and integrity of everything they submit. NSF reviewers may not upload any proposal or review content to non-approved generative AI tools.

The Volkswagen Foundation currently treats the use of generative models in or for submitted proposals as neither positive nor negative in its decisions. Reviewers may not use generative models to prepare reviews, and documents provided for review must not be entered into generative models.

Wenner-Gren strongly encourages applicants to disclose whether and how they used generative AI, in a confidential section of the proposal that reviewers cannot see. Applicants are responsible for the integrity of their content, and plagiarized writing or fabricated information leads to removal from consideration. Reviewers may not upload applications, feedback drafts, notes, or ratings to any generative AI tool.

On its Arc Artist Fellowship page, 4Culture says applicants who use AI tools to help write their narratives are encouraged to edit and personalize them so the application is unique to them. Panelists for this fellowship may not use AI summary tools to review applications or generative AI to score them or write comments.

ARPA-H has no agency-wide rule on applicants using AI; it sets rules per solicitation (ISO). Several 2026 ISOs, including HEARING, STOMP, and AIR, tell proposers not to generate solution summary or proposal text with generative AI. ARPA-H says it may itself use AI to organize and summarize proposals, will disclose this in the ISO, and keeps humans responsible for all evaluation decisions. Read your ISO's submission instructions.

For the GSS and C-FARE Policy Communications Competition (a student policy brief competition with cash prizes), entrants who use AI for any part of the policy brief must cite or acknowledge it fully and properly, following the MLA guide.

AIA New York sets AI rules per program. AI-generated images and graphics are not eligible for the Figures of Architecture competition, and AIANY Design Awards entrants must clearly mark any use of AI or AI-generated imagery in their submission.

For the Figures of Architecture design competition, AI-generated images and graphics are not eligible. No organization-wide AI rule was found.

For the Rural Health Transformation Program (RHTP), the Fall 2026 applicant guide encourages applicants to write responses from their own knowledge and experience rather than relying on AI-generated language. Applicants who use AI to help draft or edit should review responses carefully so they accurately reflect their organization.

For the Helen H. Glaser Student Essay Award, generative AI may be used only for limited editorial help such as grammar, spelling, or organization, not for substantive narrative content, reflections, interpretations, or storytelling. Authors must include an AI statement on the cover page describing whether and how AI was used.

The CORE Grants Program AI policy does not ban AI in applications but puts full responsibility for accurate, original content on the applicant and warns that unverified AI content can hurt review. For reviewers, the CORE policy cites the NIH rule that bars peer reviewers from using generative AI to analyze applications or write critiques.

For the AAAS Mass Media Science & Engineering Fellowship, applicants must affirm that AI-generated text was not used to create the submission or its figures, images, or graphics. Limited editing with tools like Grammarly is acceptable but must be declared in the application.

For the ACCP Clinical Pharmacy Research Challenge, Letter of Intent guidance lists an acknowledgement section disclosing any use of AI. Teams must describe any tool used to generate text or images, and state clearly if no AI was used.

For the Eco Concrete Student Competition, generative AI may be used mainly for brainstorming, grammar, and clarity in the report, but not to generate the core discussion and analysis, and its use must be explained in a separate section. AI-generated voices, avatars, and footage are not allowed in the video, and AI used for its text, graphics, or slides must be disclosed.

In its Advancing Innovative Research (T1D) RFA, ADA allows only limited, non-substantive AI help such as grammar editing, formatting, and reference management, and bars AI from drafting or editing any scientific section. Any AI use must be disclosed on institutional letterhead signed by the PI and Research Officer.

For the Susan S. Spencer, MD Next Generation Research Grant, applicants must disclose any use of AI to generate ideas in the proposal or to help write or edit it.

For the Hand in Hand Research Grant, a joint grant from AOTF, ASHT, and AHTF run through AOTF, the AOTF policy applies: AI writing help is allowed, applicants are responsible for accuracy, and AI-generated text must be footnoted. Reviewers may not upload applications into AI systems.

For the TechGirls Exchange Program, candidates may not use any AI tool, including ChatGPT, to help with their written responses. The review panel may end the candidacy of anyone found to have used AI in the application.

For the Military Spouse Scholarship, all essays are checked with AI tools for plagiarism and AI-generated content, and applicants who rely heavily on AI will be disqualified.

For AMNH doctoral program and graduate fellowship applications, generative AI may not be used for any intellectual content, though it is acceptable for proofing grammar, spelling, and formatting.

For the Latin American/Caribbean Conservation Research Grant, applicants may use generative AI to help with English translation but should not rely on it to write the main content. The AOS AI policy bars outside AI notetakers from grant meetings and asks anyone using AI to evaluate applications to watch for algorithmic bias.

For the Grants-in-Aid for Graduate Students, using AI tools to generate the written proposal is treated as plagiarism. Applicants unsure whether a tool complies should contact the committee chair.

The ASGE 2027 Research Awards RFA allows transparent, purposeful AI use with limits: confirm the tool does not take ownership of the output, describe AI-generated figures in the caption, and avoid AI for finding or formatting references. For the AI Scholars Program, AI may edit, brainstorm, or proofread but must not generate the substance of responses.

For the Hand in Hand Research Grant, a joint grant from AOTF, ASHT, and AHTF run through AOTF, the AOTF policy applies: AI writing help is allowed, applicants are responsible for accuracy, and AI-generated text must be footnoted. Reviewers may not upload applications into AI systems.

The Ben J. Lipps Research Fellowship AI policy lets applicants use tools like ChatGPT to help write, but they must take responsibility for the content. AI-generated images are discouraged and not allowed without a clear description of the tool, and AI used in formal research design or methods should be cited.

AUA Data Research Program applicants may not use generative AI tools to draft their Letters of Intent or full applications.

For the Amgen Scholars Program, there are currently no program-wide conditions on generative AI in applications, though host institutions may set their own rules. Applicants who use AI are asked to use discernment and fully cite sources that are not their own work.

Anthropic's research grant and API credit programs (Economic Futures, AI for Science) publish no rule on applicants' AI use. The Anthropic Fellows Program is applied for through Anthropic's hiring site, which points applicants to its candidate AI guidance: create your first draft yourself, then use Claude to refine it, and complete take-home assessments without Claude unless Anthropic indicates otherwise.

For Arts Across Arkansas, a high school arts program with cash prizes, the Arkansas Arts Council does not accept submissions created directly or indirectly by or with AI.

For the AGFC Conservation Scholarship, applicants may use AI only to brainstorm before writing and to check spelling and grammar of their own writing. AI may not write, draft, or outline the essay or video script, and misuse may lead to disqualification.

For FY27 Wheaton Cultural Project Grants, applicants submit a short AI disclosure statement saying whether generative AI meaningfully shaped their application or materials; disclosure does not change the score. The FY27 ARSG application allows AI to aid the creative or scholarly process but does not accept work samples entirely generated by AI.

For the Midwest Award for Artists with Disabilities and the NEA Big Read, Arts Midwest allows AI for editing or proofreading narratives, discourages fully AI-generated narratives, and does not allow AI-generated artwork. The NEA Big Read application asks whether and how AI was used.

AAJA sets AI rules by program. JCamp applicants may not use AI to help with applications. For the AAJA Awards, work must be human-led, any AI use must be labeled and explained, and AI photos are not accepted in photojournalism categories.

For the Trailblazers Travel Award essay, AI tools such as ChatGPT may help with prose or word count but may not generate the content itself.

For the Azrieli Music Prizes, AI may be used only for language, communication, and presentation (translation, proofreading, organizing written materials) or for electronic audio demos. AI may not be used to create music or scores, which leads to disqualification, and any AI use in the application must be disclosed.

For the Tuscora Park Health & Wellness scholarship application, using AI to write the essay invalidates the application, and applicants certify they did not use AI for any part of the application or essay.

Bayer's grants and contributions portal publishes no rule on applicants' or reviewers' use of AI. The Bayer Foundation Scientific Fellowships guidelines (September 2026, for the 2027 cycle) say the required 1-minute motivation video may not contain AI-generated content.

For Beaver County Foundation scholarships, essays must be written solely by the student, and AI or other writing assistance tools are not permitted.

For the Arnold O. Beckman Postdoctoral Fellowship, the Foundation supports AI tools as supplementary resources used ethically and responsibly, but applicants must include a brief statement with their references describing any AI use. Reviewers may not upload materials into AI tools, and the Foundation does not use AI to screen or evaluate applications.

For the Boettcher Scholarship, the foundation says generative AI tools like ChatGPT can help applicants brainstorm an essay topic, build an outline, and refine their essay. It cautions applicants against relying heavily on these tools.

For the Nancy Grace Roman Space Telescope Cycle 1 Call for Proposals, run through IPAC at Caltech and following NASA policy, any proposal material derived from generative AI must be described and cited in the References. Failure to disclose generative AI use will result in disqualification. Reviewers may not upload proposal content or review materials to generative AI tools.

For postdoctoral fellowship applications to Carnegie Science's Earth and Planets Laboratory (EPL), cautious use of generative AI as a research or writing aid is allowed if applicants take responsibility for all content they submit. Applicants attest that the application gives appropriate credit to tools including generative AI where it materially contributed. EPL says it does not use AI to evaluate applications.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's chronic disease center (NCCDPHP) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's emerging and zoonotic infectious diseases center (NCEZID) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's HIV, viral hepatitis, STD, and TB center (NCHHSTP) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's injury center (NCIPC) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's Global Health Center has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

Neither CDC nor CDC's extramural research program (ERA) has an AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not reach program or service cooperative agreements. Check each NOFO for added rules.

In the 2027 guidelines for The Creative Advantage Arts Partner Roster, the City of Seattle's Office of Arts & Culture does not restrict applicants from using AI to help write applications, but requires them to acknowledge that they used AI for writing or editing. Applicants are responsible for accuracy, and AI use is not considered in review or scoring.

Coefficient Giving has no organization-wide rule on applicants using AI. Its Project Tailwind call automatically filters out submissions that its systems indicate were generated by AI. Coefficient says its own staff use AI tools in grantmaking, for example to gather information about a grant and draft writeups.

Rules depend on the opportunity. CFNOVA's FY2027 grant guidelines let applicants use AI tools to help prepare an application without disclosing routine use, but applicants must make sure the final application is accurate, original, appropriate, and reflects their organization's work. For the Plein Environmental Scholarship Fund, an application completed by generative AI will be disqualified.

On its Corazón Grants page, CFSA says it welcomes the thoughtful use of AI tools in drafting and refining grant applications. Applicants remain responsible for making sure all content is accurate, complete, and reflects their organization's work and voice, and should take care not to share confidential or sensitive information with AI tools.

In a blog post sharing tips from its scholarship reviewers, the Community Foundation for Southwest Washington says AI is not a replacement for writing your own scholarship essay, though it can be useful for specific parts of the process. It advises applicants not to use AI to write an outline or essay.

CFA's policy for scholarship applications welcomes AI tools as long as applicants remain fully responsible for the originality, accuracy, and integrity of everything they submit. Applicants must not submit AI-written essays as their own, use AI to fake documents or achievements, rely on AI to tell their personal story, or share sensitive personal data with public AI tools. Volunteer reviewers do not use AI to evaluate applications.

In its 2026 Community Grant Program application, the foundation lets applicants prepare parts of the application with AI tools. Applicants who do so attest that they thoroughly reviewed all AI-generated content and commit to its accuracy, completeness, and alignment with their goals and representations.

For ValleyCreates Capacity-Building Grants, work samples purely generated through AI are not eligible. If AI was used as a tool to support the work, applicants should include an explanation.

In its 2026 ReGRoW RFP, Cures Within Reach says it will not consider LOIs or full proposals substantially written, in whole or in part, by large language models or other generative AI. If significant AI use is detected, CWR may remove the application from funding consideration.

DSO has no office-wide AI rule, and its current Office-wide BAA does not mention AI use by proposers. Some DSO solicitations do: the DSO Pitch Day announcement (DARPA-PA-26-01) requires proposers to say on the cover sheet whether they used AI tools and which ones, and to cite AI-assisted content. DARPA says this answer is not used in evaluation. Check the cover sheet and instructions of the specific solicitation you are answering.

For the Society for the Teaching of Psychology's Excellence in Teaching Awards, the dossier cover sheet must include a disclosure statement explaining how, if at all, generative AI was used to prepare the dossier or any materials in it.

For Arts Across Arkansas, the Arkansas Arts Council's student art program with cash prizes, the council does not accept submissions created directly or indirectly by or with AI.

NOAA has no agency-wide policy on AI use in grant applications or review. One current NOAA funding notice (the Climate Program Office FY 2027 SI/ML competition) encourages applicants to say in the project narrative whether and how they used generative AI, and bars reviewers from uploading application content to generative AI tools. Check your own NOFO, since other current NOAA NOFOs say nothing on this.

The Doris Duke Foundation has no foundation-wide rule on using AI in grant applications. One call sets its own rule. The Artists Make Technology Lab (run with Mozilla Foundation) lets applicants use AI to brainstorm, clarify language or translate, but not to write full responses, and asks applicants to declare any AI use. Check each call's RFP.

For the Electrical Contracting Innovation Challenge (ECIC) student competition, final proposals are run through software that estimates how much AI was used to create them. The ECIC jury is told about AI usage and the percentage of AI-generated content.

For Erie Community Foundation scholarships that require an essay, responses should be written by the applicant and reflect their own experiences and ideas. Applicants are encouraged to complete essays without AI tools.

Rules vary by scheme. EMBO's Scientific Exchange Grants guidelines hold applicants fully responsible for originality and accuracy and ask them to disclose whether AI tools were used and for what purpose. For the Maria Leptin EMBO Science Journalism Fellowships, using AI tools to create an application is not allowed.

For the ESPID Fellowship Award, using AI tools such as ChatGPT in grant preparation is allowed. Applicants must acknowledge any AI use in their cover letter, including the tool and how it was used.

For its Social Change Fund, Fight for Sight says it is up to applicants whether to use AI such as ChatGPT, but expects them to check that all submitted information is accurate. It says it does not currently use AI when assessing applications.

For its 2025 National Sauna Week Poster Contest, a cash-prize art contest, Finlandia Foundation National did not permit AI-generated images in entries.

Rules differ by program. For Flinn Scholars, essays must be the applicant's own work and should not be written by ChatGPT, though such resources can be helpful writing partners. For the Flinn-Brown Fellowship, the foundation recognizes that applicants may use generative AI to develop and refine their application.

Formas sets applicant rules per call. The Explore 2026 call asks applicants to describe which AI tools they used, where in the application, and for what purpose, except for basic spelling or grammar checks. Formas reviewers are currently not allowed to use AI tools to assist in reviewing applications.

Fred Hutch's HutchCAN Student Program strongly discourages using AI tools to write responses and says generative AI should not be used to write application responses. Its PACE program asks applicants who need AI tools as a disability accommodation to email hutchadvance@fredhutch.org before submitting.

On its 2026-2027 Fulbright Student cover sheet, Fulbright Denmark says all applications go through a plagiarism check and its review tools can detect AI-generated text. Applicants are strongly advised to rely on their own writing skills and authentic voice.

Fulbright Finland's application checklist for its scholar and professional programs asks applicants whether they used AI in the scholarship application process and, if so, which tools or platforms and for what purposes.

There is no Fulbright-wide AI rule, and the U.S. Scholar application instructions do not mention AI. For the Visiting Scholar Program (non-U.S. scholars), the 2027-2028 application instructions warn that copying text produced by generative AI directly into your application may hurt admission or affiliation decisions at U.S. host institutions. At least one country commission sets stricter rules: the Fulbright Commission in Egypt requires disclosure of any AI use and disqualifies applicants who use AI to replace their own thinking or without disclosure.

AI rules found are specific to the Nexa Lighthouse funding opportunity (Climate x Health Innovation: Asia and the Pacific). There, applicants may use generative AI to help prepare proposals but are expected to propose original ideas and are responsible for accurate, properly referenced content. The Nexa Lighthouse partners may also use AI tools to help process and evaluate applications.

For its scholarship applications, Grand Rapids Community Foundation says AI tools can be helpful for brainstorming or editing, but a personal statement should reflect the student's own voice and perspective.

On the Making a Difference Letter of Intent form, Greenwall asks applicants to check boxes showing how, if at all, they used AI to develop the LOI. The foundation says it is collecting this to understand how bioethics researchers use AI, and the answer is not shown to reviewers or considered in review.

HRSA has no AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. Most HRSA grants are service or workforce programs it does not cover. Check each NOFO for added rules.

For Helmholtz Enterprise proposals (Field Study Fellowship and Spin-off program), applicants must submit a declaration on whether and how AI tools were used, and AI must not be the sole or primary author of the proposal. Applicants are strongly advised not to use commercial AI tools that send data to external servers.

The Henry Luce Foundation has no foundation-wide AI rule for grant applicants. For the Luce Scholars Program, applicants may use AI tools for editing only, and must certify that their personal statement and short answers are their own work.

In its Research Grant guidelines, HFSP says generative AI should only be used to improve the readability and language of an application, not to create content, and asks applicants to disclose AI use. Research Grant reviewers are asked not to use AI tools, or AI-detection programs, in their assessments.

For the Gwendolyn Brooks Youth Poetry Awards, Illinois Humanities says entries must be the young poet's original work, so AI cannot be used, and it has a no-tolerance policy for AI-generated writing.

On its AOCA (Support for cultural operators in Africa) program page, Institut français says applicants may use AI tools to improve the drafting and presentation of applications, but the application must keep its originality and unique identity.

AI rules are set by individual IFTSA student competitions. In the IFTSA and Mars product development competition, AI may help with ideas, research, outlining, and grammar, but should not generate or substantially write the final proposal, and AI-generated visuals in slides must be disclosed. In the Graduate Research Video Competition, AI may be used for research and planning but not to create any part of the final video.

IIE sets AI rules by program. For the Artist Protection Fund, using AI tools such as ChatGPT to generate, draft, paraphrase, or substantially develop application materials is strongly discouraged and may count against the application. For the Medtronic Spark Scholarship, submitting AI-generated content without meaningful personal input, or misrepresenting your experiences using AI, is not acceptable.

In its Interfaith Civic Pluralism Fellowship 2026-27 RFP, Interfaith America permits generative AI for tasks like editing, translation, organization, or refining ideas, but applications should reflect the applicant's own thinking. Applicants are responsible for accuracy and originality, and substantive AI use should be disclosed.

For the ISN Fellowship Program, applicants may use AI such as ChatGPT or Claude to help write their applications, but must do so carefully. Applicants are fully responsible for accuracy and must proofread all AI-generated material. Fraudulent use or false statements lead to rejection and possible blacklisting.

In its Peace Research Grant instructions, the IPRA Foundation says it screens all applications for generative AI, which is forbidden. Using AI such as Grammarly to correct spelling and grammar is acceptable.

The Jacobs Foundation has not published its own AI policy. Its Jacobs CIFAR Research Fellowship is a partnership with CIFAR. CIFAR's published policy lets applicants use generative AI if they fully verify and take responsibility for AI-generated content, and prohibits using public online tools to evaluate applications, but the fellowship call does not say whether these policies apply to it. Check with the program.

The Knights Templar Eye Foundation's Career-Starter Grant guidelines ask applicants to provide a statement on whether AI and AI-assisted technologies were used in preparing the grant, and in what capacity.

For the LAF Honor Scholarship in Memory of Joe Lalli, applicants must clearly identify any part of their work samples generated by AI and explain the role AI played.

In its Creative Engagement program guidelines, LMCC allows applicants to use generative AI to prepare grant applications, as long as everything submitted is accurate and truthful with nothing fabricated. Panelists may not put application materials into external generative AI tools, and perceived AI use should not affect their assessment.

For the LF Experiment 2026 call, applicants who used AI to prepare their application are asked to describe how they used it (for example conception, writing, summary, grammar, or figures) and which AI tools they used.

For its scholarship applications, Maine Community Foundation says AI can help with learning a language, revision, or proofreading, but should not be used as a copy-and-paste shortcut. Applicants are encouraged to use their own thoughts for ideas and essays, written in their own voice.

In its Grants for Artists FY 2027 guidelines, Maryland State Arts Council requires applicants to disclose whether generative AI was used to develop, create, or refine any work samples. Samples made with AI support are eligible if the artist keeps primary creative control, but samples entirely generated by AI are not accepted.

AI rules vary by program. For Grants for Creative Individuals, applicants may use AI to brainstorm for narrative questions, but work samples created entirely by AI are ineligible and any AI used in creating the work must be explained. The Andover Cultural Council requires applicants to disclose AI used to generate art, music, or other content.

Rules depend on the program. For the 21st Century Community Learning Centers grant, applicants may use AI for brainstorming and outlining but not as a ghostwriter, and AI use must be disclosed and cited. For the U.S. Senate Youth Program scholarship, applicants must affirm they did not use AI to write their essays.

For the Bridge Funding Program, applicants may use LLMs and other generative AI tools to prepare LOIs and full applications. Applicants are fully responsible for all content, including AI-produced parts, and using these tools will not affect review.

For the Ann Theodore Foundation (ATF) Breakthrough Sarcoidosis Initiative RFP, which the Milken Institute hosts, ATF supports ethical and responsible use of AI tools, and applicants are accountable for anything AI developed or edited. Because applications are confidential, neither ATF nor Milken's SPARC will use AI to analyze submissions.

The National Academies have no single AI rule for all their fellowships and grants. The applicant guide for the NRC Research Associateship Programs and the Air Force Science and Technology Fellowship says generative AI tools should be used to enhance rather than replace your own voice, and that if you use them, the Fellowships Office relies on you to review and verify all AI-supported content for factual accuracy and relevance and to avoid plagiarism and copyright infringement. We found no AI guidance for Gulf Research Program applicants or for reviewers.

The AI rule we found is for NCA's Communication Elevation Grant. If you use generative AI in any application materials, you must include a one-paragraph Generative AI Acknowledgment naming the section, nature of use, platform, and prompt.

For the NCRA CASE Student Scholarship, the essay's ideas, analysis, and final written content must be the applicant's own. AI may not be used to generate or compose all or a substantial portion of the essay.

NED's Hurford Youth Fellows and Reagan-Fascell Democracy Fellowship programs prohibit wholesale copying and pasting of AI-generated answers. Limited AI use for generating ideas, editorial guidance, and translation is acceptable if it does not overly affect your writing style or content.

For the Slingshot Challenge, the National Geographic Society recommends avoiding AI-generated content such as images, video, or script writing in video submissions. If your video is selected for final consideration, you may be asked where your content came from, including AI generators.

For the Reflections Arts Program, entries may not be created exclusively with generative tools such as AI. If digital technology is used, entrants must describe the tools and methods, including any generative elements, in 10 to 50 words. Entrants should also check whether their state PTA prohibits AI.

NRAO's observing Call for Proposals says that, starting in semester 26A, proposers who use AI tools while preparing a proposal should tick the AI Usage checkbox on the proposal's General page in the Proposal Submission Tool.

NSDAR prohibits using generative AI tools to compose any part of a DAR scholarship application. Essays and written content must be the applicant's own original work, and applications may be screened with plagiarism and AI-detection software.

For the Action Fund National Grant Program, the National Trust discourages over-reliance on AI. If you use AI, review what it gives you and adapt it to how you would express the idea. A funded project is described as one without heavy reliance on AI to draft the application.

In its 2026 Mentored Research Award RFA, NETRF lets applicants use AI tools to help write or edit the application, does not require disclosure, and holds applicants solely responsible for all content. Reviewers may not enter any part of a proposal into AI or LLM tools.

For the 2026 National Dance Project Production Grant, NEFA recognizes that generative AI tools can support accessibility needs. The application opens with an AI statement that applicants are asked to read and acknowledge.

For the Racial Justice Fund, the New Hampshire Charitable Foundation discourages using AI in the application and says it wants to hear about your work in your own voice.

Guidance varies by program. For the Healthy Food Financing Fund, AI may help organize or edit responses, but answers must be specific and accurate. Creative Industries Division grant guidance says to avoid unedited AI-generated responses in your narrative.

NIH has no single AI rule for its prize challenges. Each challenge's own rules decide. Recent NIH challenges that address AI mostly allow generative AI if you disclose the tool and how you used it, but at least one (the Complement-ARIE Reduction to Practice Challenge) says generative AI should not be used to write a submission. Check the rules of the specific challenge before you enter.

For the NC Compass Grant run by the State Library of North Carolina (part of DNCR), applicants are asked to say whether and how they used generative AI, and AI use neither helps nor hurts an application. Reviewers will not use generative AI for merit review.

For the Women's Board of Northwestern Medicine Lake Forest Hospital scholarships, the personal statement should describe your experiences in your own words, and applicants are told not to submit AI-generated responses.

For the NASA Postdoctoral Program, which ORAU administers, generative AI tools are not prohibited, but material that is not your own product must be cited. If you use AI, cite the program name, version number, date and time, and say how it was used. Failure to cite may lead to denial of the application.

OASH has no AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from OASH's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. It does not cover OASH's service grants.

For the Community Storytelling Fellowship, Oregon Humanities discourages using AI or LLM tools to draft proposals.

Rules differ by PARC program. For the Fellowships for Palestinians, proposals must be the applicant's own intellectual work and not substantially AI-generated, and undisclosed AI use may discredit the application. For the Tanya Baker-Asad Scholarship, generative AI is strongly discouraged except for translation and grammar correction.

On its Early Career Fellowship application form, the Paul Mellon Centre permits AI tools to support clarity or formatting, but applicants must ensure all content is original and accurately reflects their research and project.

For the Thomas J Lemley Award for Health Disparities, the submission letter should confirm the work is original and created by the named authors, not written using AI such as ChatGPT.

For the foundation's scholarships, applicants are expected to submit original work, and using AI tools to generate or substantially revise essay content is discouraged. Authentic essays in your own voice receive strong consideration.

Pfizer has no grant-wide rule on AI use. Some 2026 Pfizer competitive grant RFPs (for example a US ALK-positive lung cancer education RFP and a US PTCL research RFP) require applicants to disclose whether they used AI tools to prepare the application, and say failure to disclose may lead to disqualification. Most other 2026 Pfizer RFPs we read have no AI clause, so check the RFP you are answering.

For PSA's need-based membership grant (free or discounted membership), generative AI may assist with drafting or with grammar and spelling, but submissions must be original and represent your own work. Any use of AI-generated content must be clearly disclosed.

For the 2026 Progress of Ideas Grant (Mason Gaffney Digital Collection), AI or translation tools should be limited to supportive functions, not used to generate full responses or replace your own input. Applications that look heavily AI-generated may get further review, and the application asks whether and how you used AI.

For the PRSA Houston Scholarships, the use of AI is not allowed for the application.

RWJF has no foundation-wide AI policy that we could find, but at least one 2026 call (Local Data for Equitable Communities) says applicants may use generative AI tools. RWJF warns that AI text can contain errors or read as generic and inauthentic, which could hurt the application in review, and asks applicants to check that it reflects their local context and project. Check the FAQ for the specific call you are applying to.

Roche and Genentech publish no general rule on AI use in grant applications or reviews. One Roche India program, the LEAP 2.0 scholarship challenge for final-year oncology students, told applicants that AI-generated proposals are not allowed and that an AI detection tool would be used.

The Royal Academy of Engineering's Enterprise Fellowships guidance says applicants are fully responsible for their application, which must reflect their own voice and ideas; relying solely on generative AI to write it is not acceptable. Applicants must acknowledge any generative AI use, naming the tool and how it was used.

For the Presidents' Fund, applicants may use AI tools only for limited support such as spelling, grammar, readability, and formatting. AI must not be used to generate substantive content, draft or rewrite descriptions of the proposed work or impact, or produce text that is not the applicant's own.

For the Keystone Grant, Save Our Seas Foundation wants applications in your own words. Tools that check your writing (e.g., Grammarly, DeepL) are acceptable, but applications where generative tools such as ChatGPT created the text will not be accepted.

For the New York Life Golden Futures Scholarship (sponsored by New York Life and administered by Scholarship America), submissions must not use any AI or large language model.

The Creative Advantage Arts Partner Roster guidelines say applicants may use AI to help write applications but remain responsible for accuracy. Applicants must acknowledge any AI use for writing or editing, and this information does not affect scoring.

Several current Simons Foundation calls ask applicants, in their How to Apply instructions, to state whether and how they used generative AI to develop their proposal. This is part of individual calls' instructions, not a standalone policy. The Foundation has not published separate rules for reviewers' use of AI.

The Alfred P. Sloan Foundation says it has no formal policy on applicants using generative AI. Its 2025 Metascience and AI Postdoctoral Fellowship call said using generative AI to prepare funding applications is permitted but that caution should be applied, and it pointed to UKRI's policy as useful guidance. Sloan's general grant proposal guidelines do not mention AI.

For the Latino Museum Futures Internship, applications should not rely on AI to generate answers or create portfolio materials. The program prefers personal reflection and demonstrated skills over polished writing.

For the Smith Fellows Program, applicants get a chance to disclose whether and how they used AI tools such as ChatGPT, Grammarly, or Claude. These disclosures are not penalized and are not shown to reviewers. Program staff may use AI tools for tasks like summarizing reviewer feedback and matching reviewers, but AI is not used to evaluate or score proposals.

For the John Doctor Education Prize, a short science communication video competition, all content should be user generated and use of AI is not allowed.

At least two current SOA Research Institute requests for proposals say the proposal should not be written by AI but in the researcher's own words, and any use of AI in generating the proposal should be fully disclosed. At least one SOA call for analysis also says submissions must not materially rely on AI-generated content unless disclosed or limited to editorial help.

SCBWI sets AI rules per event. The 2025 New England Spring Conference Post-Conference Writing Contest requires stories to be written entirely by the author with no AI-generated content, and the 2027 Winter Conference Portfolio Showcase for Illustrators does not accept artwork created in any part using generative AI.

For the SPJ/JEA High School Essay Contest, run with the Journalism Education Association, each essay must be original and written by the student with minimal guidance from others, including AI tools. Essays found to be written by AI, in whole or in part, are disqualified.

For the Southern Prize for Literature, any use of generative AI must be disclosed in the application, including its purpose. Using generative AI does not change the requirement that the submitted work be original and the author's sole intellectual and creative property.

For the Voice of Hope (VOH) Scholarship, AI tools such as ChatGPT, Grammarly AI, and other text-generating software are strictly prohibited, and the foundation uses AI detection software. Applications that violate the policy are disqualified without notice. Spell-check and grammar tools that do not create content are permitted.

STScI's applicant AI rules found are in the JWST Cycle 5 call for proposals. Proposers may use generative AI tools, but any material derived from them must be described and cited, and failure to disclose results in disqualification. Some uses, such as hallucinated references, are not acceptable. JWST reviewers may not use generative AI tools for any part of their reviews.

For the Alaska Department of Health's Rural Health Transformation Program (RHTP) Fall 2026 applications, applicants are encouraged to write responses from their own knowledge and experience rather than relying on AI-generated language. Applicants who use AI tools to help draft or edit should review responses carefully to make sure they accurately reflect their organization.

In its Artist Roster application, the Mississippi Arts Commission (MAC) says a human applicant must develop, review, and enter all application data, and creative work must reflect human authorship. Applicants must certify whether generative AI was used and explain how, and undisclosed use makes the application ineligible. If generative AI was used in artwork, the human-authored draft and the final work must be uploaded together as one work sample.

For the State Library of North Carolina's NC Compass Grant, applicants are asked to indicate whether and how generative AI was used to develop their application, and this will neither help nor hurt the application. Reviewers will not use generative AI for merit review.

For the United States Senate Youth Program (USSYP) scholarship, the North Dakota Department of Public Instruction's application addendum requires applicants to certify that all written components are their own original work and were not generated, in whole or in part, using AI tools or services.

The Ohio Arts Council (OAC), a State of Ohio agency, lets grant applicants use AI to help write or edit proposal text, as long as the content is accurate and represents the applicant's own work and intent. Applicants are asked to disclose any third-party AI use in an application or its attachments. OAC's policy also bars AI content from being the sole or determinative basis for any grant eligibility, scoring, or award recommendation.

Rules vary by agency. The South Dakota Department of Transportation (SDDOT) requires research proposals to describe how AI tools were used to prepare the proposal and how AI is expected to be used in the research.

For its general scholarships, the Sturgis Area Community Foundation expects applicants to write their own original essays without help from AI programs. Applicants suspected or found to have used AI to complete their essay will be disqualified.

SAMHSA has no AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. Most SAMHSA grants are service programs it does not cover. Check each NOFO for added rules.

SAMHSA has no AI rule of its own on using AI to write or review grant applications. The only guidance that reaches its applicants comes from HHS's Office of Research Integrity (ORI), which recommends that you identify any generative AI tools used in preparing a grant application and explain how you used them. ORI's guidance is a nonbinding recommendation, and it covers applications for PHS-funded biomedical or behavioral research only. Disclosed AI use that departs from accepted practice in your field may still be considered evidence of research misconduct. Most SAMHSA grants are service programs it does not cover. Check each NOFO for added rules.

Rules vary by program. The 2027 Development Track asks applicants who used generative AI in their application or submission materials to describe how and to what extent. The Sundance Collab x TikTok Micro-Series Challenge does not allow generative AI to create, develop, or revise submissions.

For the Academy of Korean Studies' 2026 contest for Korea-related teaching and learning materials, entries in the Visual Educational Content category may be created using digital tools, including generative AI.

For its scholarships, The Gerber Foundation says applicants may use AI tools to help write personal statements only as an assistant, not as the author. The foundation does not use AI or automated screening to review, sort, or evaluate scholarship applications.

The John A. Hartford Foundation's New Partners Program call for proposals lists an artificial intelligence use acknowledgment among the required proposal appendices.

For the EACR-Mark Foundation Postdoctoral Fellowships, run by the European Association for Cancer Research (EACR), any use of AI in preparing the application should be disclosed, except for simple translation or grammar checks. Applicants should list each AI tool, the exact prompts, whether AI influenced key arguments, and how they verified AI output. The disclosure does not affect assessment.

For scholarship applications, when asked whether AI tools may be used, the Philadelphia Foundation says applicants should make sure everything they submit, including essays and short answers, is truthful and written from their own unique perspective.

AI rules come from competitions ACLS runs with the foundation. For the Robert H. N. Ho Family Foundation Early Career competition in Buddhist Studies, applicants may use AI tools such as Claude or ChatGPT to write their application, must ensure content is factual and properly sourced, and are encouraged to say where and how they used AI. For the Global Buddhism Series, authors must disclose AI use in their manuscript under Yale University Press guidelines.

For the TechGirls Program, candidates may not use any AI tool, including ChatGPT, to help with their written responses. The review panel may end the candidacy of anyone found to have used AI tools in the application.

The Fulbright Foreign Language Teaching Assistant (FLTA) application's plagiarism agreement warns that some US host institutions may scan applications for AI-generated language, and that copying content produced by generative AI directly into the application may hurt admission or affiliation decisions.

Under UNCF's general official contest rules, contestants warrant that their submitted materials are created solely by them and not by AI or with AI contributions. UNCF may deem ineligible any submission containing material created by AI.

For the UCAR Next Generation Fellowship, applicants may use AI tools to brainstorm ideas for their personal statement or essays and to improve grammar, spelling, or clarity. They should not submit entirely AI-generated content as their own without meaningful revision. The program team may use AI detection tools to evaluate application materials.

The Natural Hazards Center at CU Boulder has a generative AI policy that covers applicants to its Research Award Programs. It says generative AI should never be used to create data, original scholarly writing, or related outputs, and authors who used generative AI at any point should disclose each use in the body of the submission and in an author note. Submissions that do not follow the disclosure policy are returned without review.

For VCCF scholarship applications, applicants who choose to use AI tools should use them only to check spelling, grammar, and clarity, and the ideas and examples in their essays should be their own. VCCF may deny an application based on suspected AI use in essays.

For VentureWell's E-Team Program, applicants may use AI tools to help prepare their application as long as the ideas, evidence, and plan are the team's own, and they must say how they used AI. Submitting an application that AI largely wrote is not allowed, even if edited afterward. Reviewers may not enter applications into AI tools or use AI to score, rank, or summarize them.

The W. M. Keck Foundation's Research Program application asks applicants, if applicable, to describe in up to 300 characters how generative AI was used to prepare the proposal; applications are not evaluated on this answer. The foundation may use AI tools to analyze proposals and generate summaries, and its staff, reviewers, and consultants may not share applicant data with AI tools that train on it.

Rules depend on the grant program. For the Job Skills Program, colleges may use AI for clarity, organization, and editing, are discouraged but not prohibited from using it to generate substantive content, and must make sure all facts and data are human-verified; SBCTC will not use AI to score applications. The Hospital Employee Education and Training (HEET) program asks applicants to describe the extent to which generative AI was used to develop and write the proposal.

For the Wyoming Teacher of the Year application, using generative AI to polish grammar, tighten formatting, or structure an outline is acceptable, but all applicants must honestly disclose their level of AI use in the application. The department says AI can help with the mechanics of writing but should not be the soul of the application.

For the Community Grant Cycle, Yampa Valley Community Foundation does not forbid AI, but human oversight is required to make sure all content is accurate. Applicants who use AI tools at any stage are encouraged to disclose it so reviewers are not guessing.

AFAR has not published a rule on applicants using AI to write applications. Its 2027 letter of intent instructions for the Junior Faculty and Sagol Network GerOmic awards say AFAR may use AI-assisted software from outside providers to help process applications, that application materials are not used to train AI models, and that AFAR staff and reviewers make all eligibility and funding decisions.

AstraZeneca US says it uses generative AI for initial screening of grant applications to its Medical Education Grants Office. No guidance for applicants' own AI use was found.

The Department of Justice's Office on Violence Against Women (OVW) prohibits peer reviewers from using AI to review applications. Reviewers may not upload any part of an application into an AI platform or large language model such as ChatGPT.

We found no published rule on how applicants may use AI for USAMRAA-managed funding (CDMRP announcements, now issued by the Defense Health Agency Contracting Activity). For CDMRP peer review, a September 2026 Defense Health Agency contract document (a draft statement of work for the peer review contractor) says reviewers may not use generative AI to evaluate applications and may not send application content to AI tools without government approval.

The Focused Ultrasound Foundation's applicant guide says the Foundation may enter application materials into approved generative AI systems, where data is protected by a service agreement and the model is not trained on it. The Foundation also offers an optional AI tool that gives notes on a letter of intent; this is not a review, score, or funding decision.

AIRC does not allow reviewers to upload any part of an application or its materials into online software where confidentiality is not ensured, including generative AI tools that use large language models such as ChatGPT.

FFAR prohibits peer reviewers from using AI tools in any part of peer review. Reviewers may not input proposal materials into AI tools, use AI to draft or refine critiques, or rely on AI-generated analyses or summaries for scoring, ranking, or funding recommendations.

The Foundation for Prader-Willi Research's advocate reviewer guidelines prohibit using AI, large language models, or similar tools such as ChatGPT to assist with proposal reviews.

The Great Lakes Fishery Commission says AI, large language models, and other generative AI technologies cannot and should not be used by peer reviewers of research proposals, because doing so violates confidentiality. No guidance for applicants was found.

IMLS prohibits its peer reviewers from using AI tools to analyze and critique grant applications, because it cannot control where submitted information goes. IMLS has not published guidance on applicants using AI to write applications.

The John Templeton Foundation has not published a rule on whether applicants may use AI to write proposals. Its privacy policy says it may process applicants' personal information with automated technologies, including AI, for purposes that include reviewing and evaluating funding applications, and may use that information to train those tools. It says it does not make significant decisions based solely on automated processing.

The LEO Foundation says it does not currently use AI, machine learning, or algorithmic decision-making in handling applications, and all grant applications are assessed by people. No guidance for applicants was found.

MacArthur has not published a rule on whether applicants may use AI to write proposals. Its staff AI policy says staff may not use AI for grantmaking decisions or evaluation, and staff may not put confidential grantee information into AI tools except as the policy allows. During due diligence, staff may use AI tools to gather public information, then review and verify it.

MassCEC prohibits external reviewers of its Tech-to-Market grant programs from using generative AI in their reviews, including uploading proposals to AI tools or using AI to summarize proposals or write evaluations. No funder-wide guidance for grant applicants was found.

We found no published rule on how applicants may use AI for Military Health System research programs managed by CDMRP. For CDMRP peer review, a September 2026 Defense Health Agency contract document (a draft statement of work for the peer review contractor) says reviewers may not use generative AI to evaluate applications and may not send application content to AI tools without government approval.

NIFA prohibits the use of generative AI tools during the evaluation of grant applications, to safeguard the integrity of its peer review process. NIFA has not published guidance on applicants using AI to write applications; its application guide and current funding notices do not address it.

For the Capital Development Program Fund, the Department may screen submitted applications with AI to flag responses that could make an application ineligible. Personal identifying information is removed or redacted first. We found no rule for applicants.

For the Learning Research Network (LRN) program, Rainwater may use secure AI-assisted tools to organize, analyze, and support scientific review of applications. Any AI analysis is advisory, and human reviewers make final funding decisions. We found no rule for applicants.

The Research Council of Norway tells referees they cannot use artificial intelligence, such as ChatGPT, when assessing applications. We found no rule for applicants.

Schmidt Sciences has not published rules on whether applicants may use AI to prepare proposals. Its AI Use and Guidance statement says Schmidt Sciences and its reviewers will not upload confidential proposals to public AI tools that may retain or train on them, and that Schmidt Sciences may use enterprise-grade, privacy-preserving AI tools to help administer and evaluate proposals.

The Branco Weiss Fellowship does not allow reviewers to use generative AI to analyse proposals or write reviews, or to enter documents provided for review into generative AI tools. The cited guidelines set no rule for applicants.

For the Cities Deliver: AI Solutions for Public Services call, some information from applications and interviews is given to AI tools to summarize and help review it. These tools assist human reviewers and do not replace human judgment.

Tiny Blue Dot Foundation's privacy policy says it may use automated systems, including AI and machine learning tools, to help intake, triage, evaluate, process, and report on grant applications. It sets no rules for applicants' own use of AI.

By applying, applicants agree that Wellcome Leap may use third-party AI systems to help process, summarize, or assess proposal materials, and the results may be used in its evaluation. It sets no rules for applicants' own use of AI.

We found no published guidance from Abdul Latif Jameel Poverty Action Lab on applicants' or reviewers' use of AI in grant applications.

We found no published AcademyHealth guidance on AI use by applicants or reviewers for its scholarships, fellowships, or grants.

OFVPS has published no guidance on using AI to prepare or review grant applications, and its FY2026 funding notice says nothing on the subject.

ACL has published no guidance on using AI to prepare or review grant applications. Its Caregiver AI Prize Challenge is about AI tools as the subject of entries, not about using AI to write them. ACL is not a Public Health Service agency, so HHS's research-integrity guidance on AI does not apply to it.

We found no published guidance from the Alzheimer's Drug Discovery Foundation on using AI to prepare applications or to review them.

We found no published guidance from the American Academy of Neurology on using AI to prepare research program applications or to review them.

We found no published guidance from AADOCR on using AI to prepare award, fellowship, or grant applications, or to review them.

We found no published guidance from Arnold Ventures on whether grant applicants or reviewers may use AI or generative AI tools.

We found no published guidance from Banff Centre for Arts and Creativity on applicants' or reviewers' use of AI in grant applications.

We found no published State Department or DRL guidance on AI use by applicants or reviewers. DRL's 2026 Proposal Submission Instructions, recent NOFOs and NOFO FAQs do not mention it. Check the specific NOFO and ask the listed DRL contact.

Neither the Bureau of Justice Assistance nor its parent, the Office of Justice Programs, has published guidance on applicants or peer reviewers using AI for grant applications. Check the specific NOFO for any program rules.

The Burroughs Wellcome Fund has not published guidance on applicants' or reviewers' use of generative AI in grant applications.

We found no published Campus Compact guidance on AI use by applicants, nominators, or reviewers for its fellowships, awards, or grants.

We found no published Army guidance on AI use by applicants or reviewers for awards posted under Army Materiel Command (for example the DEVCOM Army Research Laboratory / Army Research Office BAA). The BAA and recent FY26 ARL announcements do not mention it. Check the specific announcement and ask the listed contact.

We found no published guidance from Elevance Health Foundation on applicants' or reviewers' use of AI in grant applications.

EPA has not published guidance on whether grant applicants or reviewers may use generative AI. Its grant competition FAQ (December 2025) says there is no guidance on using AI when applying for an EPA grant, and its solicitation clauses and funding notices do not address it.

Neither the Foreign Agricultural Service nor USDA has published guidance on AI use by grant applicants or reviewers that we could find. FAS's most recent funding notice does not address it.

We found no published guidance from Gilead Sciences on applicants' or reviewers' use of AI in grant applications.

We found no published guidance from the Infectious Diseases Society of America or the IDSA Foundation on using AI to prepare grant applications or to review them.

We found no published IHS guidance on AI use by applicants or reviewers for its grants or fellowships.

We found no published guidance from Institute for Pure and Applied Mathematics on applicants' or reviewers' use of AI in grant applications.

We found no published guidance from Johnson & Johnson (Janssen) on applicants' or reviewers' use of AI in grant applications.

We found no published guidance from Klingenstein Philanthropies on using AI to prepare applications or to review them.

We found no published guidance from Mayo Clinic on applicants' or reviewers' use of AI in grant applications.

The McKnight Foundation has not published guidance on applicants' or reviewers' use of generative AI in grant applications.

We found no published guidance from Microsoft Corporation on applicants' or reviewers' use of AI in grant applications.

We found no published NBER guidance on AI use by applicants or reviewers for its fellowships, grants, or calls for proposals. NBER's AI page covers researchers' handling of restricted data in AI tools, not applications.

We found no published NCTE guidance on AI use by applicants or reviewers for its Research Foundation grants or other grants and awards.

We found no published guidance from National Fish and Wildlife Foundation on applicants' or reviewers' use of AI in grant applications.

Neither the National Institute of Justice nor its parent, the Office of Justice Programs, has published guidance on applicants or peer reviewers using AI for grant applications. Check the specific NOFO for any program rules.

We found no published ONDCP guidance on AI use by applicants or reviewers. ONDCP's recent funding notices do not mention it. Drug-Free Communities awards are run by CDC on ONDCP's behalf, so CDC and HHS rules for those notices may apply instead.

We found no published guidance from the PhRMA Foundation on using AI to prepare applications or to review them.

Research Corporation for Science Advancement has not published guidance on whether applicants may use generative AI to prepare proposals, or on reviewers' use of AI.

We found no published guidance from the Russell Sage Foundation on whether grant applicants or reviewers may use AI or generative AI tools.

The Teagle Foundation has not published guidance on applicants' or reviewers' use of generative AI in concept papers or proposals.

The William Penn Foundation has not published guidance on applicants' or reviewers' use of generative AI in grant applications.

We found no published guidance from the William T. Grant Foundation on whether grant applicants or reviewers may use AI or generative AI tools.