Empowering Impact: The Perils and Promise of AI in the Nonprofit Sector

This is a guest post written by consultant and IC² Board Member Jo Carcedo. Currently working as a consultant affiliate of Spring Street Exchange in Boston, Carcedo recently retired from her position as Vice President for Grants for the Episcopal Health Foundation. In this role, she developed the foundation’s grantmaking system, oversaw all operational aspects of the organization’s grantmaking portfolios, and directed the foundation’s $35 million grants budget.  

 

 

Artificial Intelligence (AI) is reshaping various sectors, and the nonprofit sector is no exception as nonprofit organizations begin to explore AI’s potential to enhance their operations, increase their impact, and address societal challenges more effectively. However, widespread adoption of AI technology by nonprofit organizations has yet to occur. In “The State of Artificial Intelligence in the Nonprofit Sector: Ethical Considerations”, authors Jared Sheehan and Nathan Chappell found that 52% of nonprofit practitioners are scared of AI. What belies the fear?

 

THE PERILS

Lack of shared knowledge and definitions

Is it artificial intelligence, machine learning, or large language models? We use these terms interchangeably, but they mean different things. Artificial intelligence is the ability of a computer to perform tasks that normally require human intelligence (e.g., recognizing patterns). Machine learning (ML) is an application of AI that allows machines to learn from data without being explicitly programmed (e.g., facial recognition), while large language models (LLM) are deep learning models that are pre-trained on vast amounts of data, think ChatGPT.

Regardless of the definition, nonprofits do not have the experience or technologists on staff to help decipher the language of AI and its usefulness therein, leaving its understanding to conjecture and a distaste for its application. Problematically, other for-profit sectors are driving current AI technology development that does not explicitly aim to achieve the goals of the nonprofit sector.

 

Bias and Trust

Much has been written about bias — an anomaly in the output of machine learning algorithms, due to prejudiced assumptions (by humans) made during the algorithm development process or prejudices in the training data. And rightly so due to an AI’s system capability to reinforce harmful stereotypes. Notable examples include Amazon’s algorithm that discriminated against women in its hiring practices and COMPAS (Correctional Offender Management Profiling for Alternative Sanctions) that misclassified twice as many black defendants as higher risk compared to white defendants. If biases are built into the data, then biases will be the output and decisions therein can yield responses and analyses that are skewed and harmful.

Can AI ever be completely unbiased? Technically yes, if conscious and unconscious assumptions on race, gender, and other ideological concepts are eliminated from the data upon which AI is trained. The National Institute of Standards and Technology of the U.S. Department of Commerce recommends a “socio-technical” approach to mitigating bias in AI. This multi-disciplinary approach recognizes that AI operates in a larger social context that must be taken into account — as purely technical approaches to solve this this problem are limited

Further, in their landmark study, “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification”, authors Joy Buolamwini and Timnit Gebru  identify increasing diversity in AI research and development teams and stress the need for inclusive datasets to mitigate these biases effectively.

If AI is to be adopted widely in the nonprofit sector, the problem of AI bias must be addressed as it is of paramount importance given that nonprofits enjoy a greater level of trust from their constituents than most other sectors — trust that can easily erode if their decisions are premised on skewed or biased data. Moreover, AI has neither the connection with the people served by the nonprofit nor the ability to understand the deeply personal narrative stories that are foundational to the nonprofit’s mission, programming and fundraising efforts.

 

Data Privacy and Security

Control and ownership of data are critical issues. Nonprofits collect and process data on clients, staff, donors, internal operations, communities and other domains. The demands for data by funding sources, cybersecurity threats and emerging regulations are forcing nonprofits to build systems to address these factors when they may not always have the capacity or expertise to do so.

 

Resource Strain

Many nonprofits operate with limited financial and technical resources, making the adoption of AI a significant challenge. The financial and technical investment required to adopt AI technology can be prohibitive, particularly for smaller organizations. Additionally, there may be a lack of technical expertise within the organization to implement and manage AI effectively. This creates a digital divide within the nonprofit sector, where well-funded organizations can harness AI’s benefits while others are left behind, exacerbating existing inequalities.

Given these significant challenges, why should nonprofit organizations embrace AI? The case for AI in the nonprofit sector is not just compelling, it is imperative. Here’s why.

 

THE PROMISE

Enhanced Efficiency and Productivity

AI can significantly enhance the efficiency and productivity of nonprofit organizations. By automating repetitive tasks such as data entry, updating donor profiles, documenting grant applications, and streamlining the reporting process, AI can free up human resources to focus on more strategic and impactful work and prevent burnout and reduce costs which can be redirected towards mission-critical activities. This shift in productivity allows nonprofits to achieve more with less and maximize their impact.

 

Improved Data Analysis and Insights

Nonprofits often deal with vast amounts of data, from donor information to program outcomes. AI-driven analytics tools can help these organizations make sense of these data, uncovering patterns and insights that would be difficult to identify manually. This can lead to more informed decision making, better targeting of resources, identification of emerging community needs, and enhanced program effectiveness.

 

Enhanced Fundraising

Fundraising is the lifeblood of any nonprofit, yet it remains one of the most challenging aspects of operations. AI can revolutionize fundraising efforts through predictive analytics, identifying potential donors, and tailoring personalized outreach strategies. By analyzing past donor behavior, AI can predict who is likely to donate, the best time to approach them, and the most effective messaging. This targeted approach not only increases the chances of securing donations but also builds stronger relationships with donors.

 

Scalability of Programs

AI enables nonprofits to scale their programs more efficiently. For instance, AI-powered platforms can facilitate remote learning and telemedicine, expanding the reach of educational and healthcare services to underserved communities. AI can also aid in disaster response by quickly analyzing data to assess needs and allocate resources effectively.

 

KEY TAKEAWAYS

Nonprofits are increasingly recognizing the potential of AI to transform their operations and enhance their impact. The opportunities presented by AI, including improved efficiency, data-driven insights, enhanced fundraising, and scalability, can significantly benefit the sector. However, these opportunities come with challenges that must be carefully managed, including ethical considerations, resource constraints, data quality and bias issues, and resistance to change.

Considering the ubiquitous nature of AI in our culture, nonprofit organizations are not in a position to sit back and allow AI systems to be developed without their inclusion-oriented input. If under-resourced, nonprofits must begin to demand that their philanthropic partners fund their internal capacity, learning, and operational infrastructures to participate in these systems.

But, first, an ethical AI framework must be developed. That framework must be based on principles of responsibility, transparency, fairness, accountability — and it must address issues of security, representation, compensation and consent, communication, power and ownership. With such a framework in place, nonprofit organizations can develop trust in these systems and learn to embrace them.

 

Related Content

Learn about AI initiatives at the IC² Institute

Read author Jo Carcedo’s full bio