AI In Philanthropy Series #4: Evaluating Impact, Responsible AI & What Funders Can Do

---

Evaluating AI’s Impact Within Philanthropy
Just as measuring impact is critical to determining the success of a grant or charitable program, evaluating the performance of AI tools used in philanthropy is essential as well. Doing so doesn’t require complex technical metrics. Rather, funders and nonprofits are asking practical questions:

• Is the technology producing accurate and reliable information?
• Can routine administrative tasks be accomplished more quickly?
• Can we free up more time for staff to focus on higher-value work like strategy and programming?
• Is the technology introducing unintended bias or errors?
• Is our staff reviewing and validating AI-generated outputs?
• Are there clear ways to correct mistakes or address concerns?
• Are we making greater impact with our work?

Starting with small pilot projects is often helpful. Testing AI in limited, lower-risk areas allows organizations to learn what works before expanding its use.

Consistently reviewing AI-assisted processes helps ensure alignment with mission and values. It also enables organizations to assess accuracy, fairness and effectiveness, and adjust practices as needed.

Helping Others Ensure Responsible AI
Philanthropists, with their resources, influence and commitment to affecting long-term change, are uniquely positioned to ensure that AI is used responsibly, ethically and inclusively. They can lead the charge in adopting the technology carefully and helping others to follow suit.

In its framework for responsible AI use, TAG states: “Responsible adoption of AI is human-centered adoption. “As society wrestles with the increasing prevalence of AI tools, grantmakers are no outlier in this evolving landscape. The role of charitable organizations is central to determining a future that balances the potential of AI to mitigate inequity and open opportunity with the potential harms to humanity. Grantmakers have a responsibility to ensure that the future of AI aligns with their core values.”

In a similar vein, CEP states: “While the for-profit sector is lightyears ahead of its nonprofit counterpart, readily adopting AI into their operations and tech stacks, the social sector is lagging woefully — but unsurprisingly — behind, given the resources needed to embark on such a transformation. Recognizing the need for a fundamental cultural shift, philanthropy has a responsibility to encourage, fund, and accelerate the adoption of responsible AI in and for the social sector.”

Here are some ways that donors and funders may be able to help others adopt AI responsibly:

1. Promote Ethical and Inclusive AI Development

• Fund research that explores the ethical implications of AI. This includes developing guidelines for fairness, transparency and accountability in AI design.
• Support AI ethical organizations. Provide grants or investments to organizations focused on building AI technologies that promote human rights and social good. For instance, in 2023 the Patrick J. McGovern Foundation awarded $66.4 million to 148 organizations working to develop platforms and applications that will demonstrate how AI can help solve challenging problems.
• Promote balanced access to AI. Fund AI projects in developing countries, or underfunded communities within wealthier countries, to reduce the risk of creating a “digital divide” where only certain people benefit from AI advancements.
• Fund research on mitigating AI bias. Support efforts to identify and lessen bias in AI models and algorithms.

2. Foster Public Awareness and Education

• Support AI education. Fund initiatives that provide AI education and training for marginalized communities, underrepresented groups in technology as well as the general public.
• Promote AI safety. Fund campaigns that educate the general public about potential AI risks and how to avoid them.

3. Fund AI regulation and governance

• Support advocacy for AI policy. Fund efforts that push for robust AI regulations at local, national and international levels that address data privacy, transparency and preventing misuse of AI.
• Promote global cooperation. Support international collaborations between countries, universities and tech companies that help standardize AI governance and share best practices.

4. Mitigate job displacement

• Support retraining efforts. Fund programs that help workers displaced by AI technologies reskill in other fields. Industries most susceptible to displacement include knowledge work, office support, customer service, financial and legal services, graphic design, writing and journalism, and software development.

5. Combat Misinformation and Malicious Uses of AI

• Support misinformation research. Fund research into tackling AI-generated misinformation, deepfakes and other forms of harmful content, and support media organizations and AI tools that promote fact-based journalism.
• Promote digital literacy. Fund programs that teach individuals how to critically evaluate information and recognize misleading content.

6. Support AI Safety and Security Research
With the growing concerns around the potential misuse of AI, philanthropy can:

• Fund AI safety research. Support initiatives that focus on ensuring that AI systems are safe and reliable.
• Promote transparent AI development. Fund efforts to ensure that AI systems are developed transparently and with a focus on accountability, such as open-source AI projects and transparency reporting by tech companies.

Given the myriad questions and concerns surrounding AI, philanthropists can help to build and maintain public trust in these innovative technologies.

What Funders Can Do in the Next 12 Months
To move from experimenting with AI to using it more intentionally, philanthropic leaders can take a few practical steps:

1. Identify where AI tools are already being used within their organizations, even informally.
2. Create or formalize simple guidelines for how AI should be used within their organizations.
3. Start with low-risk tasks where AI can assist staff, and test those uses before expanding further.
4. Provide staff with basic training on how to use AI responsibly and review its outputs carefully.
5. Talk with grantees about how AI is affecting their work and whether it is influencing funding relationships.

Taking steady, thoughtful steps allows organizations to embrace the innovation while still maintaining oversight and accountability.

It’s an exciting and transformative time for philanthropy as funders and nonprofits alike realize the incredible potential of AI to help them operate more strategically, equitably and impactfully. The potential keeps increasing, too, with agentic AI on the rise (advanced AI that autonomously reasons, plans, and makes decisions to achieve complex goals and acts upon them). As philanthropists adopt this ever-evolving technology, we encourage them to proceed thoughtfully, acknowledging governance responsibilities, addressing legal and data considerations, closing knowledge gaps, and evaluating AI’s performance. In doing so, they will be best positioned to maximize their effectiveness.

We hope you’ve found this multi-part series on AI In Philanthropy insightful, and we’ll be continuing to monitor this space closely as the technology continues to evolve.

Want to learn more about the ways we support the philanthropic sector?
Schedule a call or reach us at 800-839-0054. Together, let’s #begiving.