Can you use AI to help write a grant application? We checked what 7,954 grant-making charities in England and Wales say about AI to find out. Nearly every funder that addresses it says yes, but most attach conditions, and a growing number want you to tell them.
Most funders still say nothing at all. The ones that have spoken, though, are remarkably consistent, and their rules are a sensible guide even when the funder you are applying to is silent. Here is what they say, in their own words, with a checklist you can run through before you submit.
The short answer: Yes (with conditions)
Of the 7,954 charities we looked at, just 105 make clear whether applicants may use AI. Of those 105, 98 accept applications written with help from AI in some form. Only 7 rule it out, and most of those are prizes, scholarships or essay competitions, where the writing itself is what is being judged.
- Allowed, with advice rather than firm rules10
- Allowed, with conditions (check it, keep your own voice, protect data)62
- Allowed, but you must say you used it26
- Not allowed7
Funders that make clear whether applicants may use AI, out of 7,954 grant-making charities in England and Wales. Percentages are rounded. Hinchilla research, October 2026.
The typical statement reads something like this one, from the Coalfields Regeneration Trust's funding application guidance:
"You can use AI tools to help write your funding application; we will not reject an application just because AI was used."
Almost the same sentence appears on the websites of Forever Manchester, Somerset Community Foundation, the Clothworkers' Foundation, Northamptonshire Community Foundation and others. Funders are clearly reading each other's statements. The wording may well have started with the National Lottery Community Fund, which uses the same line, and Forces in Mind Trust says in its position statement that it is following the Fund's lead.
Two groups stand out. Community foundations and local Giving funds make up at least 22 of the 105, and 21 of them allow AI with conditions. They will not turn you down for using it, but they warn that AI-assisted applications often read as generic. Medical research charities are more likely to require a declaration, and many point to a joint statement agreed by UK research funders rather than writing their own (more on that below).
Some funders see AI as a leveller
Several funders say outright that AI can help applicants who have always been at a disadvantage in written applications. Quartet Community Foundation's guidance for grant seekers says:
"We value how AI can increase efficiency, assist applicants who are disabled or who face additional barriers to written applications, among its other helpful uses."
The Allen Lane Foundation recognises that generative AI "can help applicants that are new to applying or do not have English as a first language", and Salford CVS reassures applicants that minor spelling and grammar mistakes will not count against them.
Most funders have not said anything yet
105 out of 7,954 is about 1.3%. Plenty more funders mention AI somewhere: 783 of the charities we looked at talk about it in their annual reports or on their websites. But most of that is about funding AI research, general commentary, or how the charity is using AI in its own office. Very little of it tells applicants what to do.
Our count is a floor, not a ceiling. We searched annual reports and the grants and how-to-apply pages we had saved, not every page of every website, so some statements will have slipped past us. And the number is rising quickly: most of the statements we could date were written or updated in 2025 or 2026.
If the funder you are applying to says nothing, that is normal. But silence is not permission to paste their questions into ChatGPT and submit whatever comes back. People will still read your application, looking for evidence that you know your community and can deliver. The conditions below are a sensible default for any funder, and if you are unsure, ask them.
The conditions funders attach
The rules are strikingly similar from one funder to the next. Six come up again and again.
1. Keep your own voice
Funders want AI to help you write, not to write for you. The Henry Smith Foundation's AI statement puts it simply:
"If you choose to use AI we recommend using it to support your thinking or structure, not to write the application for you."
Young Westminster Foundation says much the same in its Brighter Futures Fund guidance: "You can use AI tools to support your funding application, not write it for you." H&F Giving is "happy for you to use AI tools" as long as "your answers stay in your own voice and reflect your organisation's genuine knowledge and experience."
City Bridge Foundation's how-to-apply page lists what to avoid, including "Submitting fully AI-generated or unedited responses" and "Relying entirely on AI to write your application". The research funders' joint statement rules out generating a whole application, or sections of one, without human involvement.
A note on where we stand: Hinchilla does draft grant applications, but it is built for collaboration. It works from your charity's own documents so the draft sounds like you, and a person should always stay in charge and decide what gets sent.
2. Check every fact and figure
You are responsible for what you submit, whoever or whatever drafted it. The Armed Forces Covenant Fund Trust says "You can use AI to help draft your answers, but you must check, edit, and take responsibility for anything it produces", and warns that "We may reject applications if we have concerns about accuracy." Barking & Dagenham Giving is blunter: "it would not be acceptable to blame AI for misleading us."
Budgets get special mention. The Road Safety Trust's AI policy says "We do not recommend using AI to develop your project expenditure", Art Fund and Northamptonshire Community Foundation warn against relying on AI for budgets and numbers, and Foundation Derbyshire warns that AI can misrepresent financial information. Build (or at least carefully review) your budget yourself.
3. Never let AI invent anything
AI tools are good at producing plausible detail that is not true. City Bridge Foundation explicitly rules out "Using AI to create false data, quotes or stories". Suffolk Community Foundation asks applicants never to use AI to create or alter quotes, and research funders such as the Meningitis Research Foundation bar using it to generate citations. If you cannot trace a case study, statistic or reference back to a real source, take it out.
4. Keep personal data out of public AI tools
This is one of the most common conditions. The Norman Family Trust's guidance for applicants is specific:
"Do not input personal data about staff, volunteers, trustees, beneficiaries, or service users into AI tools."
The Virgin Money Foundation goes further: "Don't enter personal data or data about your organisation into AI tools." Community Foundation North East's rule of thumb is a good one: don't put anything into a public AI tool that you would not be happy to publish. For a more nuanced take on how to stay safe, see our GDPR guide to ChatGPT for charities.
5. Think about the environmental cost
Of the 59 linked AI statements we read, 19 mention AI's energy or environmental impact. That is no real surprise. What is telling is that they treat it as a reason to use AI with care rather than a reason to ban it, which perhaps reflects an acceptance that AI is here to stay. Community Foundation North East discourages energy-intensive uses such as generating images and video, and Forever Manchester asks applicants to use AI sparingly. If a funder has an environmental focus, expect this to matter more to them.
6. Be careful with AI images
The Biochemical Society's guidance says applicants "must not use AI tools in generating images or videos for inclusion in their application", and Quartet Community Foundation strongly advises against AI-generated images of children or vulnerable adults. If you need pictures, use real ones, with consent.
The biggest risk: sounding like everyone else
The warning funders give most often is about quality: AI-written applications tend to sound the same, and funder after funder says they are easy to spot. The Community Foundation for Merseyside accepts AI-assisted applications but cautions:
"AI-generated applications can sometimes be quite generic and may not capture the unique voice and strengths of your organisation."
The Henry Smith Foundation is more direct: "AI-generated emails and applications are often very generic, can include errors, and don't always make sense." Sound and Music says in its application guidance: "We are noticing an increasing number of applications which appear to have been heavily edited or generated using AI tools such as ChatGPT."
Being allowed to use AI is not the same as being funded. The Clothworkers' Foundation's FAQ makes the distinction clearly: "we will not reject an application just because AI was used. However, AI-generated applications may disadvantage your chances of being funded". The Portal Trust lists, among the common reasons applications are turned down, that "Your application relied heavily on AI" and did not match what the applicant's own website said. And Greater Manchester Environment Trust "strongly advise[s] against using it heavily", because heavily AI-written applications make it harder to believe the project is realistic.
The Egmont Trust's annual report explains why this matters from the funder's side:
"The use of artificial intelligence in applications has also increased, improving presentation but making it more important than ever to assess organisational quality through direct engagement and due diligence."
In other words, polish is cheap now. What gets funded is specific: your community, your evidence, your numbers, your track record. That is what generic AI cannot supply.

Want AI help that sounds like your charity, not like everyone else?
Hinchilla drafts grant applications from your own documents, past bids and impact reports, so funders hear your voice and your evidence.
Should you tell the funder you used AI?
Increasingly, yes. 26 of the 105 funders require you to declare AI use, and several others ask you to without making it a rule. Often it is just a tick box: the Armed Forces Covenant Fund Trust, for example, says "If you use AI, please tick the box on the application form to tell us".
Declarations are most common among medical and scientific research funders. Many follow the Research Funders Policy Group's joint statement on generative AI, hosted by Wellcome and adopted by charities including Rosetrees, the Cystic Fibrosis Trust and the Royal Society. It says that "As best practice, any outputs from generative AI tools in funding applications should be acknowledged." Cancer Research UK builds the question into its application form: "You will be directly asked to declare whether you have used any generative AI tools when completing the application form."
How much detail?
It varies. The Vivensa Foundation's FAQ says "You must disclose any use of AI tools in the application form, including the tool used and its purpose." Yorkshire Cancer Research asks for the tool, the model version and what it contributed. The Norman Family Trust suggests something as short as "Some sections drafted with AI and reviewed internally".
Will declaring count against you?
Most funders that ask say no. Ataxia UK's guidance for researchers says "Disclosing this information will not impact the assessment process". Ufi VocTech Trust's policy says "The use of AI does not affect the assessment process nor the decision to award either grant funding or investment." RNID uses the answer for monitoring only and does not show it to reviewers.
Not declaring is the real risk. Music for All's policy says "Failure to declare AI use may result in your application being disqualified", and the Biochemical Society reserves the right to withdraw an application if AI use is not disclosed accurately.
At the other end, the Waterloo Foundation's statement says "You do not need to tell us if you used AI to help draft text", unless AI is central to the project itself. Suffolk Community Foundation's guidance sits in between: "We do not require applicants to disclose AI use, but we expect openness where its use materially affects the content or delivery of a proposal."
Our advice: keep a one-line note with every application saying which tool you used and what for. Then you can answer the question quickly and honestly whenever it comes up, and we expect it to come up more and more.
When the answer is no
Seven funders rule AI out, and most of them are judging the writing itself. Lincoln's Inn's scholarship FAQ allows spelling, punctuation and grammar checks, and nothing else:
"Candidates are not permitted to use AI software to write, rewrite, rephrase and/or paraphrase any part of any answer on the application form."
The others are:
- Gray's Inn, whose essay prize rules say candidates who present AI-generated work as their own will be disqualified and reported to the Bar Standards Board.
- The Nine Dots Prize (run by the KDS Foundation): "We do not accept entries that have used generative artificial intelligence (AI), Large Language Models (LLMs) and related tools as part of their creation."
- Prospect Burma, for scholarship essays: "You must write the essays yourself".
- The Royal College of Emergency Medicine, for its research essay entries.
- The Council for At-Risk Academics, for its Palestine programme, where evidence of AI use leads to rejection.
- The Jerusalem Trust, which asks schools not to use AI-generated content when describing how resources will improve learning.
A few more funders draw the line short of a ban. Humanimal Trust limits AI to "assistive tasks only". And watch out for live conversations: BBC Children in Need's guidance says "you must not use any AI tools during safeguarding or assessment calls", which rules out recording and transcription bots, and the Vivensa Foundation bans AI tools during interviews. Ask before you let a note-taking app join the call.
Why funders are paying attention
The short version: volume. The Steel Charitable Trust received 1,387 new applications in what its latest annual report calls a pilot year, and could support only 38 of them: a success rate of 3%. The report offers an explanation:
"One explanation could be that rapid advances in AI technology are making it easier for applicants to find potential sources of funding than previously."
Its trustees decided to move to invitation only for 2026-27 while they reconsider. They are not alone in connecting the two. The Allen Lane Foundation's annual report puts its rising application numbers down partly to "the increased use of Large Language Models (AI) in sourcing funders and writing applications", and the Ashley Family Foundation links a steady rise in applications partly to "the rising use of AI among applicant organisations".
When a funder closes its doors to unsolicited applications, everyone loses, including the charities that never sent a generic bid. Fewer, better-targeted applications help everyone. Our look at changing grant success rates has more on how competition has shifted.
How funders are using AI themselves
It is not only applicants. 12 of the charities we looked at say they use, or may use, AI to help assess, triage or manage applications or grant reports. Several of them use AI features built into grants platforms such as Plinth. Culham St Gabriel's Trust explains:
"AI tools within Plinth help by summarising applications, highlighting key information, and assisting with due diligence checks, enabling our team and trustees to review applications more efficiently. However, AI does not make funding decisions."
Cornwall Community Foundation says it "may use artificial intelligence to help with due diligence checks", UK Youth's grants system uses AI to help categorise data and manage its workflow, and the Jack Petchey Foundation has begun exploring AI to triage grant reports. Young Hammersmith & Fulham adds an important line: "No application is rejected or funded based solely on AI-generated information."
34 funders say the opposite: that they do not use AI to assess applications, or that no decision is made by AI. Sussex Community Foundation says "We read every application carefully, without using AI." The London Community Foundation says its team does not "use AI software anywhere in the grant making process" and will tell stakeholders if that changes. The research funders' joint statement bars peer reviewers from putting applications into AI tools, while allowing funders to use data-protected enterprise AI tools to process them.
Some funders check for AI
Kaleidoscope Trust's January 2025 call for proposals said "All proposals will be screened through an AI detector regardless." Alzheimer's Research UK "will not penalise for AI use if the original thoughts are your own", but says it will be screening applications for AI use. Camden Giving flags to its panel any application that is "overwhelmingly AI". AI detectors are far from reliable, which is one more reason to be upfront when a funder asks, and to make sure the substance of your application is unmistakably yours.
A checklist before you submit
Pulling the funders' rules together, here is what to check before you press submit:
- Look for the funder's AI statement. It is usually linked from the grants page, the FAQs or the site footer. If there is not one, use the rules below as your default.
- Declare it if asked, naming the tool and what you used it for. Keep a one-line note with every application so this is quick.
- Write the substance yourself. Your need, your project, your community and your evidence. Use AI for structure, editing and clarity (our guide to using ChatGPT and Claude for grant writing shows how).
- Check every fact, figure, name and date. Build your budget yourself.
- Never let AI invent quotes, case studies, statistics or references.
- Keep personal and sensitive data out of free, public AI tools.
- Make sure your application matches your website and annual report. Funders check.
- Read it aloud. If it doesn't sound like you, or any other charity could have sent it, rewrite it.
- Ask before using AI note-takers on calls or in interviews.
- For prizes, essays and scholarships, assume AI is not allowed unless the rules say otherwise.
Most funders that have written about AI accept it. What they want from an application has not changed: an honest, specific account of your work that they can trust. Use AI to get there faster, check everything it gives you, and tell them when they ask.
About this research
While building our funder directory, we have collected the annual reports and the grants and how-to-apply pages of thousands of grant-making charities registered in England and Wales. Between 22 and 31 October 2026 we searched that text, covering 7,954 charities, for any mention of AI, and used AI to help us read and sort every match. Where a funder linked to a separate AI statement or policy, we read that too where we could: 59 of the 65. The figures in this article cover the 105 funders that make clear what applicants may do. We left out 16 more that mention AI without a clear position on applicants using it, or that link to a statement we could not read. Every quote in this article is taken word for word from the funder's own website, guidance or annual report.
Some limits are worth knowing:
- We only searched the pages we had saved, so the real number of funders with an AI statement is higher than 105.
- It covers charities registered in England and Wales. Public and lottery funders that are not registered charities, such as the National Lottery Community Fund, are not counted.
- Policies change quickly. Always check a funder's current guidance before you apply.
Drafts that start from your evidence
Hinchilla works from your past bids, impact reports and other documents, keeps your data private and leaves the final word with you.
Starts from your own documents, not the internet
You stay in charge of every draft
Your data never trains AI models
Funder research built in
