Three ways to ethically use AI in Philanthropy right now

The hype surrounding AI right now is, in many ways, just that. However, the attention brought to Generative AI in the last twelve months and the ensuing race between Microsoft, Google, and OpenAI (the makers of ChatGPT) has resulted in numerous useful tools coming to market. The Philanthropic sector is following a cautious and considered approach, aware of ethical considerations given that AI is trained on datasets representing decades of human bias and discrimination. However, by taking a thoughtful and measured approach, philanthropic organisations can utilise AI to enhance responsiveness to the communities we support. This article explores three ethical ways to leverage AI in philanthropy right now.

One: Writing support

We do a lot of writing in the philanthropic sector, and many of us probably think we’re pretty good at it. However, even experienced writers can get stuck in particular habits, go on autopilot when writing their tenth board report, and encounter writer’s block. This is where AI can support our mahi.

AI tools such as Grammarly go beyond spelling and grammar checks, offering suggestions to reword sentences and whole paragraphs to improve clarity. Generative AI systems such as ChatGPT, Microsoft’s Bing, and Google’s Bard take it further. Copy a paragraph you’re stuck on into any of these tools, and it will restructure it for you. If you don’t like the result, ask it to try again, and provide specific feedback on what you don’t like about its result. You don’t need to use the AI’s result as-is, especially as they default to US spelling, but I am in the habit of using its suggestions to help me reorder sentences and paragraphs for clarity and punch. Regular use of ChatGPT has noticeably increased the speed with which I write and help break ingrained habits such as using a passive voice.

Two: Chatbot on your website

Applicants hoping to get a grant often visit a philanthropic website, seeking information about funding, insights on what to apply for, and help on how to apply. No matter how well we philanthropists structure our website, the difference between the words we use and the questions fund seekers ask can hinder them from readily finding the information they need. AI-powered chatbots help with this by using deeper levels of language association and processing to know that, for example, a user typing in ‘who have you funded before’ should be given information from the previous grantees’ page. In another example, a user who asks the chatbot ‘what files do I upload’ needs answers about the documents to submit for a request—associations that are obvious to us as humans but difficult to include with traditional website search.

I prefer to think of AI Chatbots as more of an enhanced type of search rather than as ‘using AI’ or a new and different tool. Chatbots have the advantage that, should they not provide a satisfactory answer, the user can send a question to the organisation for a response. Additionally, we gain the context of what a fund seeker has been looking for on the website with the direct query. There has been a proliferation of services that offer chatbots in 2023 that are easy to set up, integrate with an existing website, and offer users answers based on the current information on your website.

Three: Allow people to post in their applications once again

Grant applications have progressed from paper-based to online systems over the last twenty years, vastly more convenient for us and most fund seekers. However, some of those who could most benefit from philanthropic funding have the least access to technology. Cost limitations, limited mobile coverage, and limited technical skills mean paper-based systems will remain the most accessible option for some.

AI-assisted text and handwriting recognition have reached impressive levels of accuracy. As a result, minimal additional administrative effort would be needed to reintroduce postal submissions for grant applications. Whether handwritten or typed, basic scanners and AI can extract text for easy integration into existing funding systems. By re-including postal submissions, philanthropic organisations can ensure greater inclusivity and accessibility for fund seekers.

Conclusion

The philanthropy sector’s cautious approach to Generative AI is valid, considering the risks of perpetuating human biases. However, through thoughtful and targeted usage of AI to assist human processes, the sector can gain efficiency and accessibility that benefits our communities. Gaining familiarity with AI is important to deepen our understanding of this emerging technology and its potential impacts, good and bad.

Leave a Comment