Artificial Intelligence and Philanthropy: Where We Are

November 10, 2020

Reading Time: 4 min

AI: between hype and reality

Artificial intelligence (AI) is one of the topics of the decade. McKinsey boldly predicts that AI will add 13 trillion to the global GDP by 2030. On the other hand, Nick Bostrom of Oxford University predicts that it will become self-aware and kill us all. AI now drives cars almost as well as a human (but not quite), plays chess much better than a human, and excels in pattern recognition tasks, with very practical, life saving applications. It has also started creating art. Much podcast time and digital ink is being spent discussing when machines will be able to feel and whether that will affect their ability to do maths. Melanie Mitchell, arguably one of the more sober voices out there, thinks that there are definitely limits to what AI can do and that its promises and threats are overhyped.

AI is the mother of all trends. What are its implications for philanthropy?

Big data and philanthropy – get the basics right first

Meaningful data had great value even before big data and AI existed as terms, and it continues to have it for nearly all nonprofits and foundations. Much of this value can be readily reaped by low-tech solutions. Technology is best thought of as a tool for, roughly, making deductions (big data) and predictions (machine learning) where humans don’t have the bandwidth.

Data collection and interpretation is only meaningful with a business question. Where will rain fall tomorrow? Which kids on the list will need after-school support? How much money should I ask for? In the absence of this, the sector has been going through the motions – studies suggest that 90% of organizations collect some data and only five percent of them use it in all their decisions. (The for-profit world is of course no better – up to 85% of big data and AI initiatives fail.)

Machine learning and philanthropy – the landscape

The smart leader should therefore be prudent. Resources invested in hype cannot be used for improving operations. But not all is hype, of course. In a recent publication, Beth Kanter and Allison Fine identify ways in which AI is already being used across the industry:

Donor Matching & Personalization Engines: Machine learning and donor data gathered from explicit or implicit behavioral data to match donors with a nonprofit or cause to support. Nonprofit data is also gathered and categorized through algorithms	Facilitates more engagement through personalized communication.	• Workplace donors • Everyday donors	Philanthropy Cloud DonorsChoose GlobalGiving Ant Financial/ AliPay
Philanthropic Advising: Uses machine learning and algorithms to provide recommendations for philanthropic investment. Also includes the potential for automated impact ratings.	Helps donors make investments that yield the highest impact or strategic system change.	Program officers • Online donors	Candid Charity Navigator ImpactMatters
Donor Prediction Models and Automated Stewardship Workflow: Machine learning core methods to train algorithms on donor data to identify most likely donor prospects. Also predicts “about to be lapsed” donors who need re-engagement. Some models append nonprofit CRM data with third-party data sets.	Saves fundraisers time by automating tasks including suggesting cultivation strategies, generating draft communications, and scheduling in-person meetings with donors in person.	Major donors • Mid-range donors • Peer-to-peer • Lapsed donor	• Blackbaud • boodle.AI • Gravyty • Neon One
Online Fundraising Campaigns: Machine learning analytics to analyze donor databases, sometimes appended to third-party data sets or social media data. Also includes the use of chatbots.	Used to personalize donor engagement & communication, convert donors via customized landing pages and analyze unstructured social media data to personalize donor communication at scale.	Everyday Donors	Nonprofit Cloud • Quilt.Ai • CivisAnalytics • Persado • Chatbot platforms • GiveLively
Donor Research/Data Collaboratives: Platforms that are sharing data for research purposes with privacy protocols and ethical standards.	Data is used for research to better understand giving patterns	All donors	• GivingTuesday • Fundraising Effectiveness Project
Reporting and Workflow Tools: Uses machine learning and natural language processing.	Helps platforms efficiently generate reports or automate administrative tasks like customer service for causes & donors, reports, and website content	• All donors	• GlobalGiving • DonorsChoose • Crisis Text Line • USA for UNHCR • GiveLively

Source: AI4Giving, 2020

The Altruist League and artificial intelligence

Our investment advisory decisions are increasingly aided by our machine learning algorithms; our artificial intelligence (AI) platform is now firmly at the core of our value proposition. For the moment, the AI tool assists our analysts and investment managers as they keep track of movements, preselect them for the Index, and advise members in a B2B format. Our goal is to open the platform up for B2C use by mid-late 2021 so that individuals, too, can use the advisor to build their own custom philanthropic portfolios and invest directly into systemic change around the world.