This post was originally featured on VentureBeat
It’s no secret that philanthropy is an imperfect way to fund important causes. When we give with our hearts instead of our heads, we tend to overfund popular organizations and underfund needier ones, creating a discrepancy in charitable giving. For instance, the diseases we donate to and those that kill us are incredibly misaligned, and reflexive giving strategies often end up funding dysfunctional organizations instead of effective ones.
With wealth concentrated in the hands of an increasingly narrow minority, causes valued by top earners (and their foundations) also naturally receive the most charitable funds. Because of this, wealthy donors have a disproportionate responsibility to look beyond personal bias and allocate their funds wisely.
Is there a better way? With artificial intelligence on the rise, there could be.
As Aristotle once stated, “to give away money is an easy matter and in any man’s power. But to decide to whom 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.”
As focus on effective altruism grows, the famous philosopher’s statement rings truer than ever. To give is simple, but to give effectively based on evidence and reason requires rigorous research, willpower, and money management. Which charities use their funds most responsibly? Which will do the most good in the least amount of time (if that’s your goal)? Which align best with your values while also benefiting the greatest number of people?
These are quantitative questions, yet givers have been historically motivated by feelings of goodwill and empathy (not to mention tax incentives), which jettison statistics for sentiment. Psychologist and Yale professor Paul Bloom goes as far as to argue that empathy makes the world worse by prompting us to care more about a baby that’s fallen into a well than, say, global warming.
This is a human problem that could be balanced out with some perspective (and well-applied technology). AI could supply the cold eye to match warm hearts, prevent irrational decisions, and even out the charitable playing field.
How? Machines, unlike people, can crunch numbers quickly and give objective advice that’s entirely unhampered by emotion or impulse. As Rhodri Davies, head of policy for the Charities Aid Foundation (CAF) explains, “An AI with access to vast quantities of data and the ability to analyse it at greater depth and speed than a human ever could is going to add value when it comes to identifying most acute pressure points in terms of social or environmental needs at any given time.”
Davies goes on to outline how AI could identify the most pressing needs, provide a framework for effective giving, align with donors’ values, and maximize donor satisfaction. It could also reduce the cost of philanthropic advice by opening it up to the mass market.
AI chatbots already offer advice across a variety of industries, and philanthropy is no different. Arthritis Research UK, for instance, uses a virtual assistantdeveloped with IBM Watson to give site visitors tailored information about the condition. Such a chatbot could, in theory, provide personalized advice to donors about where and how their funds can be put to use.
Facebook, whose AI investments and tailored news feeds make frequent headlines, has its own philanthropic feature that could theoretically match users with charities based on preference, much as it does with brands. Mark Zuckerberg seems primed to head in that direction, especially considering that his charitable organization CZI (Chan Zuckerberg Initiative) acquired an AI startup called Meta. The acquired tech helps scientists navigate, read, and prioritize the millions of academic papers in existence. Such a tool could do the same for charities or donors seeking to effectively parse information and allocate resources.
As with other industries, involving AI in philanthropy is not without potential problems. Do the algorithms reinforce biases? Whose definition of “good” will researchers program an algorithm to prioritize? And can we trust that the data used is accurate and complete?
This latter question is especially crucial as we begin to build solutions. While we may have enough data now, we’ve quickly moved from too little data to too much. The burden, then, is on tech developers to create tools that can really and truly access, analyze, and report on broad sets of data with enough accuracy to be useful. The whole idea of using AI to maximize giving efficiency, after all, assumes it can represent the entire spectrum of organizations to choose from, not just those with available data.
If AI can achieve this, as I think it will, advocates of effective altruism would see their vision more nearly realized. At the same time, effective altruists are also concerned with AI as an existential threat and are researching the societal risks posed by reckless AI implementation. Bottom line: We need to take steps to mitigate AI’s downsides even as we use it to advance charitable initiatives. As long as we can keep the technology in check, our planet and life on it will surely reap the benefit.