Charitable Fundraising and AI - New Report

AI and Fundraising

A report published by International Fundraising Thinktank Rogare gives some really thought provoking insights into the way charities should use AI for fundraising and those roles for which AI is at present entirely unsuitable.

Ethics

The research asked Chat GPT a series of questions about the ethics of fundraising. They noted that Chat GPT could answer each question quite neatly, listing factors to consider, but that the AI only answered the question directly, and would not give any contextual or particularly nuanced answers. For example, none of the answers dealt with donor dominance unless specifically asked about it. Donor dominance is “any serious form of undue or inappropriate influence that a donor or group of donors exerts over a nonprofit organisation or its staff”. The research concluded that to have a question answered by AI, you already needed to know about it.

So, should AI be used to resolve ethical fundraising issues? Not yet. In the future AI may develop more nuanced knowledge about fundraising ethics, but at present it doesn’t have the ability to safely make ethical fundraising decisions.

So what is AI good for in fundraising?

AI can be used to automate many aspects of fundraising:

• analysing donor data;
• identifying potential donors;
• contacting them with personalised materials; and
• soliciting donations.

There is a very real potential to undermine public trust in charities if these are not done appropriately.

Concerns raised by the researchers:

I will not reproduce the several pages of concerns, but here are a selection:

• A myriad of data issues, from ownership, to consent, to the veracity and quality of data used by AI.
• Biases – perpetuating historical biases about race, religion etc and codifying discrimination into automated processes.
• Human oversight is needed to spot misleading, false and discriminatory content generated by AI – but the fundraising sector does not currently have the required level of technological literacy,
• The roll out of AI will be inequitable, benefitting the richest charities most, widening the gap between the richest and the poorest charities.
• AI could, without substantial human supervision, enable emotional manipulation of the most vulnerable groups.
• Any lack of transparency around the use of AI, or donor mistrust of disclosed AI fundraising may both discourage donations.
• Where AI is adopted too quickly by a charity under pressure to ‘not fall behind’ problems can often arise.

Conclusion

The paper has explored the ethics and risks of using AI in fundraising and considers the use of AI to solve ethical dilemmas in fundraising. The researchers have proposed a ‘research agenda’ to guide charities in the adoption of AI in a way that meets the ethical and legal needs and expectations of donors, beneficiaries and national regulation.

For further information about the use of AI for charities, contact Neil Burton [email protected]

 You can read the full report from Rogare here