Can ChatGPT write my case for support?

by Laura Phipps and Ben Beaumont - 20 December 2023


There’s no escaping the hype around AI-driven chatbots like ChatGPT, Bing Chat and Bard. It’s hard to imagine an industry untouched by artificial intelligence of some kind but AI chatbots have added impressive capabilities into the mix.

At More, we’re following the rapid developments in the field to better understand how AI and chatbots like ChatGPT could support fundraising. Together, we've been exploring how these applications can help (or hinder) the development of fundraising messaging, specifically cases for support.

In this blog, the third in our case for support series, we’re asking ‘Can an AI chatbot like ChatGPT write my case for support?’.

The answer is, as with many things AI-related, complicated.

What are AI chatbots?
These sophisticated chatbots are underpinned by Large Language Models (LLMs) like GPT and LaMDA. The easiest way we’ve found to understand this technology is that it’s a computer programme trained on natural language. This training may come from books, articles and online content from the internet, as well as through dialogues with human testers. As a result, these models can predict the next plausible word in a sentence with impressive skill.

This is undoubtedly an oversimplification of decades of mind-blowingly complex research. But thinking about chatbots in this way can help to explain their current strengths and limitations. These models are not a ‘person’ and they’re not ‘thinking’. They’re computer models looking for the next sequence in a pattern of words – only they’re exceptionally good at it.

The result? In the words of ChatGPT itself: ‘AI chatbots have emerged as extraordinary tools that can generate human-like text, often indistinguishable from that penned by real individuals’.

How chatbots can help you build a case for support
What does all this mean for fundraisers looking to develop copy to set out a compelling case to their donors?

Here are some key observations we’ve made through our interactions with AI chatbots like ChatGPT, as well as some watch-outs.

1. They help you put something on a blank page
If you’re finding it hard to get started, AI chatbots like ChatGPT can really help. If you were a charity promoting education to disadvantaged young people, you could ask a chatbot to summarise key reasons why a donor should care about access to education or the impact of education on a young person’s future outlook. If you have streams of thought from your own internal development work, you might ask an AI chatbot to summarise them for you in a way that would inspire a donor to support. This can spark ideas and get you going with the first line.

2. They’re getting better as a research tool
When these chatbots first launched, we found it hard to source and verify any factual information they were giving back to us. That makes sense – they were developed as language models and not search engines, and some (like ChatGPT) had only been trained on data up to September 2021. This is now changing.

While their knowledge is still shaped by the data they were trained on – and the inherent biases and untruths within it – chatbots like Bing Chat are increasingly being integrated with search engine functionality. While this makes it easier to source references for factual statements they produce, there is still reason to double check what they’re telling you.

3. They (understandably) get things wrong
In searching for the next plausible sentence in a sequence, these AI-powered models can sometimes create statements that are very believable but simply not true. These are so-called ‘hallucinations’ and can sneak past even the most eagle-eyed human. In the words of one AI expert, their ‘ability to fabricate is just amazing’. They can also make cultural blunders because they don’t have the context of knowing all the sensitivities around particular issues. That’s why you remain an essential partner in the development and writing of any fundraising narrative.

4. They’re an aid and not a replacement
Only you (the human) know the strategy behind your case for support. You know your donors’ motivations and, therefore, the arguments that will connect most deeply with them. You know the institutional priorities and the difference that donations will make to the causes your organisation serves. However well-trained an AI chatbot, it does not know that detail unless you go to great lengths to feed it all into the system.

So, while you remain the best person to develop the framework for your case for support and the arguments you need to set out, chatbots like ChatGPT can be really useful writing aids.

5. They’re a copy editing partner
With the right instructions (called ‘prompts’), a chatbot can copy edit your case for support for style, clarity, length and structure, e.g.‘Re-write this paragraph to make it shorter and punchier, while still delivering the key message for donors that x, y, z’.

If you don’t like it, prompt again with more specific guidance. Doing so can challenge your thinking, spark ideas about structure and flow, and help your writing pack more punch. It can help you to personalise copy too, tailoring it to suit different types of donor.

But be careful! We’ve seen examples where copy edits can subtly change a message or statistic, affecting its accuracy. Chatbots also tend to strip out some of the seemingly superfluous words that may actually add colour and distinctiveness to copy. Remove too much of this and your case risks sounding sterile and business-like.

You have all the context for the document you’re writing, so you have an important job to check whether it’s saying the right thing in the right way for your donors.

6. It takes time and practice
It takes time to learn how to use AI chatbots most effectively. Everyone thinks and works in different ways, but there’s no doubt that a chatbot like ChatGPT can bolster your weaker areas if you harness its power in the right way. Nothing beats playing around with a real-life example to understand what it can – and can’t – do for you.

The quality of what these chatbots can produce is also really dependent on how good your prompt is. Prompt engineering may well become a key workplace skillset in an AI-driven world and it takes practice to get the best results. If you want to understand more about prompt engineering, check out free courses like this one from Coursera.

7. Be conscious of privacy
This might be our final point but it’s really important. Privacy is definitely a thing when it comes to using AI-driven chatbots, especially when developing copy that may be private or sensitive to your institution. Some corporate firms have banned the use of chatbots due to the risk (or reality) of staff inadvertently releasing sensitive information through them.

When using a chatbot, be clear how the information you’re entering may be used. Some may tell you that your data is protected, but others may require you to adjust privacy settings within the tool.

In conclusion, we think that AI-driven chatbots hold huge power for helping fundraisers craft copy that can inspire donors, and their capabilities will only grow. But at the moment, they won’t do it alone. You remain the most important person in building an argument and making a case for support that is deeply informed by your own knowledge and experience.

If you’d like some real-life human support to develop your case for support, More Partnership can help you research, articulate and test your philanthropic narrative for maximum impact. To discuss further, contact lphipps@morepartnership.com.

*This article is up-to-date at the time of publication, but we expect developments in AI chatbots to continue to move at pace.