Artificial Intelligence (AI) can be a useful ally for charities. But using AI brings responsibilities. Funders are starting to watch closely, and misusing AI, or trusting it too much, can hurt your proposals chances of success. This guide explains how your charity can use AI responsibly in grant writing while protecting your integrity and data.
Why This Matters
Funders such as the European Research Council and the European Commission are already reviewing how AI is used in proposals. Professional grant writers are expecting stricter rules. At the same time, charities are adopting AI quickly: over three-quarters now use it in some form. With grant writing taking anywhere from 40 to 200 hours depending on the size of the bid, AI can reduce workload, but only if used carefully and transparently.
Using AI in a Safe and Practical Way
AI can help you put an ok draft together, it can help with brainstorming, summarising, and structuring documents. But it should never be the final author of your proposals. Staff must always review, edit, and tailor outputs to reflect your charity’s own voice, data, and stories.
A simple record of how you’ve used AI, for example, noting where you used AI or which prompts were used. That can help you keep track of what prompts you’ve used and to what benefit. Plus you’ve a proven record of transparency if a funder asks and reassures trustees or colleagues about the process. But you don’t need to go over board, instead of logging every single word or conversation, just note:
- AI used to draft first outline of executive summary (heavily edited by staff).
- AI used to generate initial project timeline (adapted by project manager).
Above all the final text must use your organisation’s own evidence. AI-generated paragraphs should be rewritten to highlight your unique impact, using your evaluation data, case studies, and programme results. Consider your organisations data privacy rules before you give all this information to ChatGPT, Microsoft Copilot, Google Gemini, or Meta AI. Question if the AI tools you’re using for grant writing will train future models on the information your providing.
Handling Facts, References, and Security
AI can sometimes invent or misquote information, so it is vital to double-check every fact. Replace suggested references with citations from your own research, reports, or publicly verified sources.
I don’t mean to harp on about confidentiality but it is a key issue. Public AI tools are not secure enough for sensitive material. So never paste in partner letters, unpublished reports, financial data, or personal information about staff, volunteers, or beneficiaries. If your charity chooses to use AI more deeply, consider exploring private, paid, or on-premise tools that offer stronger data safeguards.
Examples of How AI Can Help
Besides the big open AI platforms like ChatGPT, there’s a whole host of other grant focussed AI powered platforms that can help you both find grants and generate proposals. By scanning thousands of funding opportunities, matching them to your organisation’s profile, and then auto-drafting application text.
Many systems allow you to upload past proposals and project documents so the AI “learns” your mission and data, then uses this context to tailor answers to a funder’s requirements. Most of them use large-language models (GPT-family or similar) but tuned with proprietary data (award-winning proposals, sector databases, or your own documents) to improve relevance and accuracy.
One big benefit is getting past the blank page stage, generating initial drafts. Provided the service has suitable security and privacy features, you input project goals, budgets, and key statistics, AI grant writing tools can write sections of a narrative, draft budgets, or propose outcomes.
Susan Mernit, a nonprofit consultant, earlier in March described using “custom GPT Research Brains” loaded with her past successful proposals and program data. So when a new RFP appears, the AI highlights alignment with funding criteria and produces a draft outline, cutting the time she spends on the proposal by roughly 30%
Here are some charity-friendly ways to put AI to work safely:
- Drafting an executive summary based on a project description, which you then expand with your own evidence and case studies.
- Turning bullet points into a clearer draft of a case study, which you then check for tone and accuracy.
- Simplifying complex sections so they are easier for non-specialist funders to understand.
- Producing a checklist of required documents based on published funder guidelines, which you then verify against the official list.
Training & Resources
To use AI effectively and ethically, charities should invest in training and resources:
- Attend webinars or workshops on AI in the nonprofit sector.
- Explore guides from organisations like NCVO, or Charity Digital.
- Create internal policies on acceptable AI use. (According to this years Charity Digital Report over three quarters of UK charities are using AI, but just 44% of charities have a digital strategy).
Conclusion
Using AI responsibly means preparing for funders’ expectations. Some may ask you to disclose how AI was used. Others may focus on originality and authenticity.
Be ready to explain your AI use clearly and honestly. Run similarity or plagiarism checks before submission to ensure originality and review drafts for bias. Provide training so staff know when and how AI is acceptable within your organisation.
