Article last updated 2 October 2025.
The implications of artificial intelligence (AI) for all aspects of society are huge, and the charity sector will not escape. However, while there are considerable fears about ethics and job replacement, AI also presents many opportunities for the sector in terms of generating income and creating efficiencies.
For example, AI tools could be used to help charities offset increased costs such as the impending rise in employer national insurance contributions (NICs), according to Andrew Maxwell, investment director for charities at Rathbones. He says: “After the budget we have seen a large number of companies, including some of the UK’s largest charities, write to the chancellor to say that under the current fiscal arrangements they will have to make redundancies. That’s a real concern. If AI can be harnessed to increase efficiency, it could reduce the pressure on charities.”
Simon Lapthorne, Rathbones senior research analyst, observes that there are areas where efficiency improvements can enable organisations to address unmet needs, especially as the sector faces constant criticism for spending too much on administration. “Individual donors have become much more aware of this and scrutinise the figures. One of the things they look at before giving money to a charity is finding out how much of it goes to the frontline. Ultimately, charities could help a lot more people with the money saved from using AI effectively.”
Investment applications
From an investment management point of view, Lapthorne opines that there are two elements to how AI could be applied to charity portfolios. “One is improving the efficiency of administration, for example in client communication, perhaps using AI to write the first draft of a report, which enables managers to spend more time on the investment side and less time doing admin.”
Then there’s the investment decision-making process. “At this stage we are not aware of anybody who is using AI to any meaningful extent to replace investment managers and analysts in making investment decisions,” he says. “But that may come in the future. For now, our human investment managers are deciding which companies to invest in for charity clients, and of course that can include those involved in or impacted by AI.”
He describes generative AI as a general-purpose technology. “Discussion of AI these days tends to revolve primarily around the technology sector. The single most important event in the equity market last year was probably the third-quarter results from Nvidia. But actually, the implications of AI are very profound and much more widespread than technology. It will have implications across all sectors. Every company has the potential to be impacted by AI, most obviously through productivity enhancements. The global annual wage bill is $44tn. If companies can save even a relatively small proportion of that through investing in AI, that justifies spending a lot of money on it.”
As an example, Lapthorne cites an industry which charity investors might not automatically assume to be driven by technology – soft drink manufacture. “We’ve got two big companies there. If one of those is using AI to do loads of productivity-enhancing admin stuff and one isn’t, it could create a competitive advantage. So it is really important for all organisations to be looking at this. They all have a marketing function, an accounting function, an IT function, a human resources function etc.
The application of AI will help improve productivity and not necessarily replace people. The history of technology has’t generally been that jobs are destroyed, but that jobs are changed so that people can be redeployed and do different things.”
There is a potential for a skills gap to open up between the skills required for what people do now and those that are needed for what they might do in the future, he warns. “But I think there is one thing that perhaps mitigates that and is different from what happened in the past with technology. Previously, the jobs that were replaced tended to be relatively low-skilled manual blue-collar type ones. However, a lot of the jobs that are going to be displaced by AI will be middle-ranking clerical jobs where it may be easier to redeploy those people. Maybe they have transferrable skills for example, and can be redeployed without significant re-training more easily than somebody who only knows a single, specific, manual function and would need more re-training. But we shall see.”
Crucially, that productivity improvement could be used in charities as well, if they deploy a little bit of AI to streamline some of their admin processes and redeploy people to the frontline. “They could spend less of their money on bean counting and more on active charitable work. I would see AI as a positive for charities who are not like other companies whose customers don’t tend to scrutinise the cost base.”
Portfolio construction
So what does this mean for portfolio construction? Maxwell is looking to create diversified portfolios that target long-term growth themes. “We have certain themes that permeate our portfolios which are commonly aligned with charity sector values – things like ageing populations, the energy transition and the digital transformation. AI can be applied to every one of these areas. It perhaps is just the timing of the impact that is still to be determined.
“When we think about companies in each of these themes, we want businesses that have strong competitive advantages and with AI, we want those best placed to capture those advantages. When you consider, for example, the internet, the winners were those companies that could most effectively use the new technology to their advantage. It wasn’t necessarily the ones who built the modems. A great example is Amazon. It started life as a bookshop. But it utilised the power of the internet better than just about every other company.
“The benefit that charities have is that they tend to have a multi-year, often multi-generational time horizon, so they can tolerate short term volatility to target longer term growth themes.”
Ethical concerns
Aside from concerns over job replacement, Maxwell is alert to a number of other ethical considerations. “We are investment specialists by profession, so that’s where we focus our efforts. We help charities grow their capital so they can deliver on their purpose. That said, we are hearing of a number of areas of impact generally in the charity sector. AI could significantly improve fundraising efforts by helping to write fundraising bids. But it doesn’t mean an end to human interaction which there will always be a place for in the charity sector.”
Of the other areas of ethical risk, data privacy is the obvious one. Maxwell says that the large language models are based on huge volumes of data, some of which is publicly available. “But the amount of data on the internet is absolutely enormous and there are concerns that using a significant proportion of it could breach data protection regulations.”
Lapthorne adds that the public internet contains an awful lot of “material the veracity of which cannot be relied upon”. And the way the models work can also create problems. “Because they are trained on the public internet going back many years, some of the material includes a lot of built-in historic biases. For example, if you ask it to give you a picture of a doctor, it will usually be a man, while a nurse will often be a woman. An investment analyst will wear glasses. And those biases can get reinforced. There’s also the question of what happens when you start training AI on material that itself was generated by AI. There’s some evidence that after a few iterations it reduces to gobbledegook. There are a number of challenges in terms of getting outputs that are accurate, ethical and lawful – although that having been said, the models are getting much more sophisticated.”
The road ahead
We could be living through an example of what is called Amara’s law, according to Lapthorne. “This states that the impact of new technologies tends to be overestimated in the short run and underestimated in the long run. “We could be heading towards a period where this massive ramp-up of spending on AI pauses. I think there’s a lot more still to come over time, but we could reach a period over the next 12-24 months where there’s some digestion; where people consolidate a little bit of what they’ve done.”
He cautions that the more fundamental threat would be if AI models don’t evolve as expected, or there’s some kind of AI-related disaster which confirms ethical concerns, or draconian regulatory interference that puts a brake on AI evolution because politicians are scared they can’t control it.
“But I think that on balance, the trend is still very much positive – although it would be dangerous to assume it all goes in a straight line and that there won’t be bumps in the road along the way.” Charities need to be prepared to travel along that road as smoothly as they can, and harness the many advantages AI offers, while remaining alert to the risks.