Shaf Mansour explores the implications of AI for charities
Changing working practices spurred on by the pandemic have accelerated the implementation of digital infrastructure for all industries. But in recent months new leaps in artificial intelligence (AI) have offered the potential for the next major revolution in how we work.
With the rise of digitisation leading to a wider scope of tech choices and less resistance for senior buy-in, charities of all sizes are beginning to take on new digital tools, but is AI ready to take centre stage in the third sector?
What can AI do?
The major new development in AI has been the rise of tools capable of plain-language interactions.
For charities, the opportunities could be game-changing. For example, you could feed an AI information from your database and ask it to ‘find a list of donors in London most likely to respond to our new summer campaign’. Using the database, and looking at past trends, previous and existing donors are quickly identified. The magic here is that the command could be made in plain English, and the AI does the rest.
Repetitive tasks, such as verifying postcodes and data to maximise Gift Aid claims could be made trivial by AI, saving time behind the scenes for those who had to manually trawl through data. But artificial intelligence can do much more than search databases and notice trends.
Programmes known as generative AI, such as OpenAI’s ChatGPT-4 and Google’s Bard have recently been making headlines for their ability to quickly generate highly specific and human-sounding passages of text.
Given the same level of plain text instructions, a generative AI can quickly draft outreach emails or content for campaigns. This could be simple or go as in-depth as ‘draft three versions of an outreach email asking for donations to a new campaign supporting animal shelters. The tone should be light and friendly and should contain the phrase we need your support at the end’.
For smaller organisations, complex day-to-day tasks can be simplified and work can be done on a scale much closer to large charities – without the need for larger teams, external agencies, or prohibitively huge budgets.
What's the catch?
The potential for AI is massive, so big in fact that we don’t know the full power of even currently released models – this is known as capability overhang.
This is indicative of how new these large AIs are, we still know very little about how they work, and they aren’t perfect.
When it comes to details, artificial intelligence can’t tell if information is necessarily true, it will simply ‘believe’ the data that it’s fed and do its best to understand what it means.
While this lack of critical comprehension often leads to trivial funny mistakes, it could easily spiral into bigger issues. Even with good background information, generative AI could make a claim from a mistaken understanding of context, which a human reader would infer.
It is every user's responsibility to ensure that these tools are treated with caution and that means those working with AI are advised to consider its work in the same way they might with a junior or entry-level member of the team. It’s a valuable tool, but everything that it creates needs to be reviewed prior to sharing more widely, especially if it’s public facing.
Similarly, charities should be cautious about copyright issues. Organisations posting large amounts of generated text could find themselves at the wrong end of a copyright claim, and with no evidence of where the ideas originated from it could become a serious issue.
It is unclear at the moment how secure AI systems are, with some reports claiming AIs can be reverse engineered to give up personal information. Anyone considering using AI with their databases should be wary of potential data protection concerns.
Sensitive information should be treated with the highest sensitivity and not shared with any external system unless authorised to do so. Those working with AI systems with any uncertainties about how data could be used should steer clear of doing so.
Innovation is a great thing, but always staying ahead of the curve can be a risky business. Organisations, charitable or not, need to find a balance between risks when engaging with AI tools. It doesn’t always hurt to sit back for a bit to monitor, learn, and see how things develop.
What’s next for charities and AI?
When making choices about the technology that charities use, they should take everyone on the digitisation journey – whether that’s AI or implementing CRM software.
A charity CRM, for example, shouldn’t be just for the data team, it’s a database that whole teams should be able to access, understand, and use to inform their work and provide more value for their supporters.
Digital tools for communicating with supporters, creating fundraising appeals, and gathering data on their donors already exist and still hold a lot of value today.
By using integration tools such as Zapier with Access Charity Websites, charities are already able to easily link up a huge number of systems, enabling them to automate tasks, such as sending personalised messages to people who make a donation of certain amounts.
It is safe to say that AI tools are going to continue developing at pace, and we are likely on the precipice of an AI industrial revolution – but this doesn’t mean everyone needs to jump on the bandwagon just yet.
Charities should embrace new systems, but start by focusing on the goals they want to achieve and working backwards to find the tools to help them achieve them. For many, investing in AI too early may be too much of a gamble, especially when the tried and tested tools available today can do the job.
Ultimately, there is still some way to go before we can wholly rely on artificial intelligence to do all the work required by charities, especially when it comes to replacing the value of the human touch. It’s here to help us, not to take over from us.
Shaf Mansour is product manager at The Access Group's not-for-profit division.
