Rhodri Davies, Programme Leader, Giving Thought

Rhodri Davies

Head of Policy

Charities Aid Foundation

The role of giving


2 June 2017

Artificial intelligence (AI) is a technology that will bring huge opportunities as well as challenges for charities and their supporters. We have looked at some of the challenges in previous posts, and in the companion piece to this blog we consider in detail one area of real opportunity that is currently untapped: namely, the provision of philanthropy advice.

In this blog, I want to give a sense of how AI is already affecting the work of charities. Some of them are using the technology in interesting ways for research or to deliver services. Others, however, are beginning to see AI ─ and in particular the need to ensure appropriate ethical safeguards – as a new charitable cause in itself. Here are five examples.


One fairly straightforward way that charities can use AI right now is by adapting one of the growing number of AI personal assistants or chatbots. For example, Arthritis Research UK have partnered with Microsoft to pilot a service based on its Watson AI that can provide users with tailored information about the condition. This is an approach that could be applies in all sorts of areas, as it essentially just involves taking information that a charity probably already provides in the form of individual webpages or fact sheets and turning them into a responsive user-friendly service. And one benefit of an AI-powered approach over using human advisers (apart from the obvious reduction in cost) is that it can make the service more accessible, as there is no danger of an engaged phone line, no need to stick to office hours etc.


AI can also help to widen access to services by overcoming language barriers. Currently, most charities do their best to ensure that their services are accessible for people who speak a different first language. This may mean getting materials translated, or employing specialist translators to help out with face-to-face meetings where required, which is often extremely expensive and presents other challenges too (e.g. it requires people to disclose often highly sensitive personal information to a third party).

The Children’s Society has begun experimenting with using Microsoft’s AI-powered live translation tools to try and overcome some of these barriers in its work with young refugees and migrants in London. This technology enables you to hold a direct conversation using a mobile phone or VOIP software (such as Skype), and have your speech translated into another language in real time. This can bring huge benefits if you are trying to provide a service that relies heavily on personal trust.


There are a number of examples of organisations using AI to tackle poaching in Africa. For instance, a team of machine learning experts at the University of Southern California has developed an AI-driven service called Protection Assistant for Wildlife Security (PAWS), which analyses years worth of data on poaching behaviour and uses game theory to suggest optimised routes for ranger patrols. This has already been tested in the field in both Uganda and Malaysia, and the evidence suggests that it does help to make patrols more efficient.

The Lindbergh Foundation in the US has developed a programme called Air Shepherd, which uses unmanned aerial drones to patrol conservation areas and record footage. They have worked with a company called Neurala to apply deep learning algorithms to the data from these drones, with the aim of teaching them how to recognise poachers. They are now putting this AI to work analysing the video feeds from the drone network in real time (including infra-red night time footage), in order to identify poachers before they reach herds of elephants and prevent them from doing so.


It was announced earlier this year than Mark Zuckerberg and Priscilla Chan’s philanthropic venture the Chan-Zuckerberg Initiative (CZI) had purchased a startup called Meta, which had developed an AI that could help scientists navigate, read and prioritize the millions of academic papers in existence. The hope is that this could make it easier for scientist to find research that is related to their own, and reveal valuable findings that might be hidden in obscure journal, and this speed up the process of scientific advancement.

This is not a directly charitable application, but it is relevant for at least two reasons. Firstly, CZI is a philanthropic vehicle (even if some have questioned this, as I explained here), so its purchase of Meta suggests that Zuckerberg and Chan see what the AI start-up does as social good and thus an appropriate focus for philanthropy. The second reason is that many charities (particularly in areas such as healthcare) do an enormous amount of research as part of trying to achieve their mission, and could benefit enormously from a tool of this kind – particularly as CZI has indicated that it plans to make the facility freely available in the future.


Quite a lot of the focus on AI and philanthropy to date (and certainly the majority of what you’ll find if you google it) has been on the potential dangers posed by AI (including everything from reinforcement of social bias to total annihilation of the human race, as famously highlighted by Professor Stephen Hawking), and the need for philanthropic funders to support initiatives designed to combat or at least mitigate these risks.

This has, perhaps unsurprisingly, become a hot topic du jour for many Silicon Valley philanthropists, and also tied in with the Effective Altruism movement (on the basis that the potential impact of an AI doomsday scenario is so great that even if the probability of it happening is vanishingly small, it still makes rational utilitarian sense to try and address it. (This is an example of the “Pascal’s Mugging” pheonomenon highlighted by philosopher Nick Bostrom, and something we have looked at before on this blog).

Concerns of this sort have led the Open Philanthropy Initiative (A joint venture between Good Ventures, the philanthropic foundation of Facebook co-founder Dustin Moskovitz and his wife Cari Tuna, and the Effective Altruism organisation GiveWell) to make “potential risks from advanced artificial intelligence” on of its focus areas.

We can see that both the opportunities and the challenges presented by AI are already beginning to influence the work of charities. There are also areas of great potential where the technology has not yet filtered through, such as in the provision of philanthropy advice (as I explore in the companion piece to this blog). What seems certain is that new applications of the technology will continue to emerge that could provide opportunities to deliver more effective interventions, and could also transform the way charities work and the ways in which people are able to support them.