Artificial intelligence in philanthropy

In a remarkably short period, AI has transitioned from tech labs to our fingertips, becoming indispensable in both our professional and personal lives. This new era of easy access to AI technology is rapidly transforming every aspect of our existence, with the intersection of artificial intelligence and philanthropy emerging both as a powerful force for good and a safeguard against potential misuse and abuse.

To delve deeper into this fascinating convergence, we sat down with Professor Giuseppe Ugazio from the University of Geneva who is a leading expert in AI and its applications in social and environmental impact. Professor Ugazio is part of a team at the University of Geneva who are investigating how philanthropic organisations can lead the ethical and inclusive AI revolution and leverage AI to enhance governance, decision-making, and impact assessment. Their research will culminate in a Handbook on Artificial Intelligence and Philanthropy, set for publication in 2024.

Our discussion delves into how AI is revolutionising philanthropic efforts, enhancing decision-making, and driving more efficient outcomes. Additionally, we examine how philanthropy is playing a crucial role in ensuring the ethical use of this transformative technology.

Interview with Professor Giuseppe Ugazio on AI and its impact on philanthropy

By Christoph Courth, Head of Philanthropy services, Pictet Wealth Management

Christoph Courth: Professor Ugazio, thank you for speaking with me. How do you see AI supporting philanthropists throughout their journey?

Giuseppe Ugazio: Thank you for speaking to me on this fascinating topic Chris. AI has the potential to support most of a philanthropist’s activities, from the simpler example of using generative language models to draft calls for grants, to the more complex tasks of identifying partners, measuring impact, or determining efficient resources allocations. While AI is being used increasingly by businesses, in philanthropy it seems only a few actors have embraced the power of this new technology.

The most fertile ground for turning to AI to get support is where there is plenty of data, and where interpreting this data would require a large amount of (skilled) human resources. AI can help make enormous amounts of data digestible, aiding the human eye to see patterns and define accurate strategies.

A concrete example is using natural language processing to process large amounts of text and extract the most significant features from it. Imagine a philanthropic organization is implementing an education programme impacting thousands of individuals. We can ask each of these individuals to provide a one-page feedback explanation on how the programme affected them – positively and/or negatively. We can then use AI to generate a summary document identifying the most common positive aspects mentioned in the reports or the most frequent negative externalities. In this way we can make a data-driven assessment of the programme, based on feedback from beneficiaries and we can use thisdata to determine whether this programme was successful, what could be improved, and if there were some negative externalities we did not anticipate. So, AI can help us overcome the core challenges of grant making, listening to and learning from our beneficiaries.

CC: Can you provide specific examples of organisations effectively using AI in their philanthropic work?

GU: The Altruist League is among my favourite organisations using AI in the philanthropic space: they use it to support their activities at all levels: to identify promising grassroot movements around the globe working to promote the social goals the League believes in; they further use AI to monitor the progress and impact of these movements, for example monitoring their coverage in the media or identifying legislation introduced to regulate the causes supported by these movements. This approach heavily relies on AI addressing two of the most important hurdles for non-profit organizations’ efficiency: 1) having to seek funds and prepare reports and 2) using AI to direct philanthropic capital to these movements, in particular for match-making between donors and individuals and/or organisations.

Another example is the Italian Fondazione AIS who are using large language models to assist in their quest to develop an impact measurement tool that can be applied to both philanthropic projects but also to sustainable finance initiatives.

More examples can be found in the handbook we at the University of Geneva are developing, which will be available open-access once it is published end of 2024.

CC: What considerations should philanthropists keep in mind when using AI?

GU: The first consideration that comes to mind is about bias and inclusion. Due to the digital divide, i.e., the fact that some populations have more digital information than others, there is the concern that societies who are less digitally advanced will be excluded from the benefits of using AI-based systems. A related concern is that recommendations provided by AI algorithms will be tailored to better suit data-richer populations. In philanthropy it could lead to less funding being directed to the less digitally rich regions.A second consideration is about being left behind and not leveraging the power of AI. In doing so philanthropist risk becoming inefficient or their work obsolete as this could lead to resource waste and/or a misallocation.

CC: In your view, how can philanthropy contribute to the development of ethical AI?

GU: At an AI and Philanthropy conference we recently hosted in Geneva, Ravit Dotan, the award winning AI ethics advisor, researcher and speaker, identified sixmain areas where philanthropy can play a role in shaping ethical AI: as grant makers, users, developers, buyers, investors, and social justice advocates. For example, as grant-makers “foundations can harness the power of AI for social good by funding non-profits that have responsible AI as part of their core product.”Christoph Courth: How do we balance AI's efficiency with the need for human judgment in philanthropy? Giuseppe Ugazio: AI should be used to complement not substitute human judgment. AI can help us get more precise and accurate information to make informed decisions, it should not however be used to make decisions for us. In philanthropy this is particularly important because whether a decision is “correct” or the best one largely depends on our own individual values and beliefs.

CC: What risks do philanthropists face if they rely too heavily on AI?

GU: The largest risk I see is the loss of diversity, of risk taking, of exploring. AI is based on algorithms that give a large premium to maximize reward by minimizing the risk of failing. This is leading to what some call the “boringification of culture” (Patrick Ryan blog), i.e. strongly limiting creativity and innovation and favouring minimal departures from “known ground”. In philanthropy this is particularly dangerous as it would push us to overinvest in low-risk projects that typically are also the ones with least potential for large impact.

CC: How do you envision AI shaping the future of philanthropy?

GU: It is very hard to say, but the hope is that AI will help provide more and better tools to help philanthropists collaborate with one another as well as with other actors, coordinating more efficiently their different initiatives. This is particularly important given the high level of complexity of the issues addressed by philanthropy, which require a systemic approach.

CC: Finally, what advice would you give to philanthropists starting to explore AI?

GU: Do your research: 1) be aware of how to properly use AI tools (for example, being mindful of what information is ok to be shared – e.g. not sharing confidential information). 2) identify what the needs are and what AI tools exist that can be used to fulfill this need. 3) be aware of what AI can’t do to avoid being disappointed by the results.

Our partners at StiftungSchweiz are developing an AI-learning journey, which will be a useful resource to help guide philanthropists.