In the past decade I worked in an executive capacity very close to a number of charities, first and foremost Lloyd’s Register Foundation, and non-profits. In fact I set up and run until recently a non-profit business, Safetytech Accelerator Limited.
Armed with these rich perspectives, recently I have been discussing the use of AI with a number or charity and non-profit leaders and advisors, and comparing their views with those of leaders in the commercial space.
While many of the AI use cases are similar to those of commercial businesses (internal efficiency, automation etc), charities and non-profit stand out in terms of where AI can make the most difference, and present the biggest challenges.
In this short article I wanted to focus on these stand-out aspects.
Solving the Impact Challenge
AI offers a powerful solution to one of the sector’s most persistent hurdles: rigorously demonstrating evidence of impact and establishing a robust Theory of Change.
It is well known that most charities and their funders / donors, are constantly facing the challenge of demonstrating the impact of their interventions both at the time of building business cases and during and after the delivery of a particular intervention.
Despite this potential, only just over 10% of charities are currently using AI for impact and evaluation reporting.
AI can help in at least two ways:
- Stress-Testing Strategy: AI can democratize complex strategic planning. Leaders can input their Theory of Change into an AI model for a “stress test,” allowing the AI to act as a sceptical social scientist to critique underlying assumptions and suggest external risks before grant money is spent.
- Quantifying Outcomes: AI can move organizations beyond simple output metrics and assist in gathering and organizing outcomes and impact data, harvest them from social media, automate feedback loops, and more.
For instance, AI could automate sentiment analysis on thousands of beneficiary feedback forms, allowing to track emotional data over time, or facilitate longitudinal tracking (e.g., automated follow-ups) to gather elusive long-term data without taxing limited staff capacity.
This is not a detail: it has been reported that charities that are using digital tools and AI to create robust, quantitative outcomes reports, achieved significant increase in donations.
As more charities adopt digital tools and AI to quantify their impact, charities that don’t may rapidly become less competitive. Will there be a tipping point?
Ensuring Ethical Engagement
The use of AI can amplify the impact of charities. However, both in service delivery and increasingly by grantees and third parties, the use of AI introduces risks that demand proactive management, especially regarding those served by the charity.
The key areas include:
- Ensuring that grantees use AI responsibly and ethically. Granting organizations will increasingly need to enforce that AI adoption by grantees aligns with charitable purpose and organizational values, and understand how to mitigate significant external risks, including reputational.
- Protecting beneficiaries: Charities must be mindful of the risks associated with AI, particularly for higher-risk beneficiaries, such as children, or those holding sensitive data. Leaders must ensure that AI applications are appropriately “guardrailed”, for example, ensuring chatbots are prevented from dispensing potentially misleading medical or legal advice.
- Bias and equity: Algorithms can entrench human biases, risking discriminatory outcomes if accountability is not prioritized. Concerns over the potential for bias and discrimination in one of the biggest barriers preventing large charities from moving forward with AI.
The Charity Commission has in fact started to look into AI, providing useful guidelines for management and Trustees (see e.g. here https://charitycommission.blog.gov.uk/2024/04/02/charities-and-artificial-intelligence/)
Turning Challenges into Opportunity
Despite the potential, and specific risks they face, many charities can encounter significant internal challenges when trying to adopt AI. These issues stem from a lack of digital and AI skills among both leaders and staff and risk aversion.
Many CEOs and trustees do not feel confident about their understanding of AI, which affects their ability to provide effective strategic guidance. Staff capacity adds to the problem, with about most charities stating they have insufficient digital expertise among their employees and volunteers. Consequently, most charities find it hard to use AI tools in their daily operations.
Yet, I strongly believe the convergence of AI and purpose-driven organizations is inevitable. There is a unique opportunity right now to rewrite how success is measured and deliver even more value. The skills gap should not deter charities, and risks can and should be managed.
By working together to adopt these tools responsibly, a new era of philanthropy that is data-driven, transparent, and profoundly human could be upon us.