Philanthropists often claim that it is easier to make money than to give it away. It’s a sentiment rooted in the challenges faced in making good decisions in grant making and it’s one that has existed for a very long time. If we go all the way back to 350BC we find a frustrated grant maker in Aristotle when he said, “to give away money is an easy matter and in any man’s power. But to decide to who to give it and how large and when, and for what purpose and how, is neither in every man’s power nor an easy matter".
It’s amazing to think that in the nearly two and a half thousand years since Aristotle’s writings so little has changed around how philanthropists feel about their giving. Since the time of Aristotle our economic, social, cultural and political existence has changed significantly, with much of human progress influenced by advances in technologies. So where is the pervasive influence of technology in traditional grant making? Outside of some notable exceptions, philanthropy looks and functions in much the same way as it did 100 years ago. It’s no surprise therefore that the frustrations around how to make good philanthropic decisions remain prevalent.
Could AI revolutionise philanthropy?
So is it overly optimistic to think that the fourth industrial revolution, with the advance of machine learning, supercomputing and autonomous vehicles, will see philanthropic decision making made easier? “Why not?”, asks Orion Matthews, a technologist, non-profit expert and entrepreneur who has worked across a number of senior roles in philanthropy in the United States. In a recently published article on LinkedIn he wrote about the potential for using Artificial Intelligence (AI) in philanthropy. With all the information at the fingertips of grant makers today, perhaps what is truly limiting us is our human capacity to process it, “sometimes the right people are missed, the right organization is rejected, or the right partnerships are never even considered.”
Of course there are real risks in believing digital technology has the cure for all that ails us in philanthropy. I recently facilitated a discussion on the role of philanthropy in democratic societies hosted by Perpetual. On the panel was Lucy Bernholz, who writes extensively on her blog Philanthropy2173 about the future of good and the role of digital technology in our civil society. I asked Lucy whether, in a world where philanthropists are often accused of making irrational decisions around their giving, there’s a future where decisions are left in the hands of an AI with a cold eye approach to the art of giving. In a post Centrelink automated debt recovery world, Lucy reminded us that the AI is born of the code that is written by the human, who in turn is flawed with their own biases and beliefs around what constitutes the most good for the most deserving.
Let’s face it, biases abound in philanthropy without the help of AI, we’re human after all. Whether it’s the unconscious bias that sees women and girls’ projects fall behind the pack in funding levels, or even the confirmation biases that keep us supporting the same types of projects and organisations year on year, humans can find a myriad ways to mess up decision making. So perhaps there is a happy medium where AI reduces our biases if we’re willing to put its design into the hands of those individuals and organisations that have most at stake – our beneficiary partners. Is the advance of AI in philanthropy actually an opportunity to democratise our decision making processes?
Philanthropy by algorithm
Over the past year I’ve been working with a family who has been experimenting with the idea of using an algorithm to help them better shortlist proposals for consideration. This family funds research and social programing and as such are rarely comparing apples with apples in their decision making process. One family member in particular has contributed an enormous amount of time into the development of the algorithm and over the past nine months has trialed and tested its efficacy in shortlisting projects submitted via Perpetual’s annual application process. Initial testing threw up challenges around interpretation on data and biases in how those actually filling in the application form perceived what the funder might consider the ‘right’ answer.
Before progressing to the second phase of their testing, the family agreed they needed actual applicants to interrogate the algorithm and application form questions behind the process. Opening up philanthropic decision making to the scrutiny of the outside world is rare and may not have been done by this family was it not for the need to ensure the integrity of the proposed approach. I gathered a group of eight senior NFP leaders, including CEOs and senior development people to provide their feedback. I must admit, I went into the meeting expecting criticism of the use of an algorithm but was surprised by the level of enthusiasm. That’s not to say there wasn’t push back in some areas, but overall those in the room felt if an algorithm meant more rational decision making in philanthropy they were for it.
I’m no digital technology evangelist; I see the advance of the machine as a two sided coin, one that will provide our social, cultural and environmental worlds with amazing opportunities underscored by terrible risks. At the coalface of opportunity and risk in our communities sits the perfect space for philanthropy. How grantmakers both use and support organisations working in a world in which digital technology pervades all will be important to our civil society.
I’m a believer that great philanthropy happens with a warm heart and a cold eye. That the human capacity for empathy that occasionally leads us to irrational decision making also leads us to want to give in the first place. Philanthropists will never find making good decisions easy because there is so much at stake and AI won’t change that, but it may just help make those tough decisions a little more impactful.