How to maintain trust when using AI in fundraising, grant seeking and grant writing

We’ve joined forces with Strategic Grants (www.strategicgrants.com.au) for a series of articles looking at AI use in fundraising, donor marketing and grant writing.

Building trust with donors is foundational to successful fundraising. A funder’s eagerness to donate is based on their personal connection to an organisation, the organisation’s track record delivering genuine change and how they can demonstrate that impact.

An over-reliance on Artificial Intelligence (AI), or undisclosed use of AI, has the potential to erode funder trust. A lack of transparency may lead funders to question whether the content reflects real organisational experience or is fabricated by AI.

AI should be used to enhance efficiency, without compromising trust. To maintain credibility, all claims should be backed by verified data, real world case studies and measurable outcomes.

If the sector experiences declining trust in AI-generated content, this could contribute to more invite-only or closed application processes, limiting equitable access.

How can organisations responsibly use AI while building and maintaining trust?

Policies, guidelines and governance

Firstly, organisations should ensure there is transparency about organisational AI policies and ethical standards. Avoid practice that tries to hide or obfuscate use of AI.

You should develop internal AI usage guidelines or custom instructions to ensure tools behave in line with organisational values, ethical expectations and your approved policies.

• Build structured review steps into workflows to identify errors, bias, or missing perspectives.
• Implement a human-in-the-loop model: always review, verify, and fact‑check all AI-generated material before use.

Before you jump into the latest and shiniest AI tools, ensure you have completed a risk analysis and have explicit approval in line with your organisation’s policies. An internal AI policy should clearly outline what tools you are permitted to use, and what can and cannot be input into those tools.

Support the human-in-the-loop

Organisations need to invest in training and support to build their staff’s capability to use AI tools effectively, and ethically. (Joining our AI community is a good place to start!)

Maintain strong relationship-based fundraising practices such as face-to-face engagement, real-time storytelling, and transparent reporting of impact.

Considered prompts

Staff should use clear and specific prompts to improve quality and minimise bias or hallucinations. It’s a good idea to actively ask the AI to provide counterpoints or test its own assumptions to strengthen rigour.

Don’t rely solely on AI

It’s good practice to supplement AI-generated content with diverse, reputable and contextually relevant sources, never rely solely on the model’s output.

Utilise AI tools trained on your content and information

Train your own AI secure tools to edit and refine your own organisational information to leverage the efficiency AI is intended to create.  This will ensure it is only drawing from your own validated data, evidence and experience and is not presenting hallucinogenic information from external, unverified sources.

Be transparent about your AI use

Continue to maintain trust with funders and donors by using AI ethically and transparently. Some funders already ask about the use of AI in preparing funding applications. Ensure you declare your use of AI when required to do so.

Understanding AI’s limitations and biases

AI can sometimes amplify biases and stereotypes

Language models can mirror cultural, gender and socioeconomic biases in their training data. Never assume what AI has provided you is balanced, as important perspectives, particularly of people less represented across the mainstream, may not have their voices heard. This is especially true for First Nations communities whose knowledge AI can easily misrepresent. To ensure ethical outcomes, AI initiatives must respect First Nations data sovereignty by incorporating traditional knowledge and recognising communal rights.

Further Reading: A Guide to First Nations Data Sovereignty (ID-SOV)

Confirmation bias risk

AI is eager to please, and can be prone to reinforcing the framing of your questions and prompts. This also means it can deliver results that seem confident when unchallenged. Challenge these responses, challenge your assumptions, and test alternative questions.

In grant-seeking, this often shows up when you prompt AI to make the case for your funding needs. It will build a confident, one-sided argument, when assessors want to see evidence, a clear grasp of the problem and honesty about risk. The same danger appears when you use AI to interpret your own evaluation data for a funder report. It can confirm the story were hoping to tell rather than what the numbers actually show, undercutting the meaningful, trust-building reporting that strengthen your case for the next grant.

Gaps in coverage

AI often favours popular or recent information. Niche, historical or diverse viewpoints may be overlooked. The best way to close that gap is with context. The more background you give on the who, what, when and why of your task, the better AI can work beyond its defaults. Just be careful what you share. Before you prompt, strip out personal and sensitive information or use a tool with proper data protections in place.

This matters most in prospect research where AI regularly returns outdated information, and sometimes even fabricated grant amounts, closing dates or eligibility criteria. You need accurate information and evidence to approach the right funders, so check every detail yourself against the funder’s own source.

AI is powerful, relationships are stronger

While AI can assist not-for-profits in their fundraising and grant seeking activities, it’s important not to use these tools at the expense of relationships with donors and potential grant givers.

Following some of the above steps will help set you up with the foundations for fundraising success.