While artificial intelligence (AI) has not reached the kind of general intelligence exhibited by fictional characters like Hal 9000 from “2001: A Space Odyssey,” the current, popular AI systems that exist – generative, large language models like ChatGPT, with which this blog was co-authored – are rapidly being deployed, experimented, and used in the philanthropic space.
Though it is always difficult to predict long-term change, especially during hype cycles in the technology sector, the exponential progress of AI over the past few years is worth paying attention to. At its best, AI can shape the future of charitable endeavors and help counteract the generosity crisis, which has seen donor retention plummet. At its worst, AI reinforces systemic challenges, demanding careful strategy. As AI continues to influence various aspects of our lives, it is helpful to begin exploring both its potential promises and its possible perils for the nonprofit fundraising sector.
Understanding AI
While other AI systems exist, for the sake of simplicity, I’ll mainly focus on large language models, like Meta and Microsoft’s Llama 2, or OpenAI’s ChatGPT. While large language models have already existed, the more recent rise of commercially available AI systems combined with public popularity and experimentation have driven widespread buzz. The building blocks of large language models are based on fine-tuned interactions with vast amounts of open data, usually sourced from the internet. Large language models use these data, deep learning methods, and advanced statistics to respond to user prompts. These prompts may be tailored based on an organization or individual’s topical focus, location, funding needs, RFP requirements, etc.
The Promises of Using AI in Nonprofits
Should nonprofits embrace AI? And what does using AI mean for fundraising? The type of AI that exists right now can be thought of as a paradigm shift in personal assistant technology. It can, through iteration and optimization, find novel perspectives simply via prompts. AI, then, may hold great promise to change the way work is conducted for nonprofits.
In fact, foundations and organizations are already racing to fund projects harnessing AI for social good. For instance, Google.org has committed $25 million in grants to support 15 nonprofit projects who are using AI to address Sustainable Development Goals (SDGs), including reducing air pollution, enhancing global food security, and improving maternal health care. Additionally, within the first program of the Vaccine Confidence Fund – a joint effort managed by Global Impact to harness the power of social media and online engagement to drive vaccine confidence – several studies by grantees employed chatbots and machine learning to assess their potential in boosting vaccine confidence through social media. These efforts yielded promising results, providing strong motivation for continued exploration of the role this technology can play in the future, especially in the second program of the Vaccine Confidence Fund.
Outside of these examples, here are four promising use cases where AI can help with nonprofit objectives and fundraising goals:
1. Strategic Engagement: AI serves as a strategic ally in optimizing engagement at scale. For example, nonprofits can create highly effective donor profiles by consolidating comprehensive information gathered during the prospect research process. After using AI to distill giving history, personal interests, business connections, etc., nonprofits can then tailor their outreach efforts, again using AI, ensuring that they are more impactful, efficient, and attuned to individual and community needs.
AI tools can also help analyze strategic donor profiles to suggest personalized engagement techniques like A/B testing for different segments of donors, thus refining outreach strategies continuously. Through AI-driven methods, organizations can closely monitor and refine their engagement metrics and conduct stakeholder interviews to enhance cultivated relationships strategically.
2. Creative Fundraising Campaigns: AI has more to offer than just high-level strategy; it may also fuel the creative engines of nonprofit professionals. With specific, context-sensitive, and continually refined prompts, AI helps craft imaginative, compelling campaigns that inspire positive action. Whether writer’s block strikes or you need an extra dose of creative inspiration, AI-generated content may elevate the quality of campaign materials. From wordsmithing to proofreading, AI helps organizations become stronger storytellers, amplifying their appeal for wider visibility and, ultimately, boosting fundraising efforts.
3. Data-Driven Operations and Decisions: AI-powered customer relationship management software and its analytic capabilities may enable nonprofits to identify the most effective interventions and optimize resources. One example can be feeding your organization’s financial data to AI and asking it to find room for budgetary improvement. By understanding data-driven approaches through AI, your staff’s decision fatigue may be exchanged for streamlined recommendations, insight to reimagine feedback and performance data, and enhanced operational efficiency. However, it is crucial to prioritize data privacy and ensure that sensitive information is appropriately safeguarded throughout these processes.
4. Programmatic Work: From to-do lists, email backlogs, and grant writing, AI can boost productivity by reshaping vital administrative functions within nonprofits. Beyond simply incorporating expert advice and industry best practices, AI can also automate various activities, freeing up time to focus on more creative or intensive efforts to drive mission-oriented objectives.
Amid a reported crisis in generosity, according to the recent GivingUSA report, AI emerges as a promising tool for increased efficiency and innovation to possibly counter this crisis. Now is the time for organizations to invest in retraining and upskilling to gain a competitive edge in an increasingly digital world. When AI is integrated strategically, nonprofits can focus on their core strengths, such as fostering curiosity, collaboration, and care for others.
The Perils of Using AI in Nonprofits
While AI is a valuable tool, it is crucial to remember that it lacks the ability to understand deeply personal, narrative stories – a domain where humans excel. Since the nonprofit sector inherently centers on connectivity and empathy, there is more urgency to ensure balance and exercise caution. As is well documented, there are legitimate concerns with AI, including increased misinformation and automating bad behaviors.
Here are four key risks to be aware of:
1. Ethical Concerns: In the case of AI, if biases are built into the data, then biases will be the output. Like humans, not only do built-in biases have the potential to do harm, but they also skew responses and analyses. Additionally, given the opacity of AI, it is complicated to fully comprehend its decision-making process. Who benefits from the outcome of a generated AI response? After all, AI is far from value neutral. AI practitioners and the data fed into AI systems may be implicitly influenced with social and political values. As AI increasingly impacts high-stakes decision making, the goal is to ensure that AI serves the betterment of society as a whole and does not perpetuate undesirable or unaligned values.
2. Abstraction and Hallucinations: AI systems are still evolving and refining. They can, however, occasionally produce ‘hallucinations,’ generating false or contradictory information in an overconfident manner as if it were true, thus making it harder for users to determine the veracity of information. An example of hallucination occurred in this JAMA Oncology research letter, where the authors used AI for cancer treatment, and it provided inaccurate information.
Despite producing text that may seem knowledgeable, hallucinations occur because AI systems often generalize based on word distribution and lack a stable, human understanding of the world, which can sometimes lead to unpredictable results. Even with ample warnings from companies that emphasize that AI outputs are not intended to offer professional advice, including legal, financial, or medical guidance, nonprofits should emphasize continued oversight and independent verification when using AI.
3. Human Dependency and Labor: AI systems, at the end of the day, still rely on human labor. This encompasses everyone from scientists in corporate labs to independent wage laborers who are tasked with labeling data for certain systems. The dependency on human workers, and their working conditions, is an important consideration for nonprofits using AI.
4. Cultural Implications: AI systems often train on data that may include pirated content. This raises both additional ethical concerns about the sourcing of data and further poses questions around authorship, copyright, and plagiarism. To say the least, it is a complex issue within the broader context of AI.
To navigate these perils, nonprofits must lead with responsibility, transparency, fairness, and accountability. Among other things, nonprofits should also prioritize privacy and data rights, and address inequalities where possible. It is imperative, then, to have staff aware of ethical AI practices to avoid worsening disparities that already exist. Staff should first ensure that AI systems align with their intended purpose and the organization’s values. It is also essential to uphold sustainable practices, data ethics and security, as well as diversity, equity, and inclusion principles. Continuous monitoring and awareness of AI’s impact is crucial, ensuring that it is used to enhance, rather than replace, the human aspect of relationships and care within the sector.
Final Takeaways
AI is not a replacement for human capabilities, but it may enable nonprofits and fundraisers to guide human dignity and flourishment via more creative ways to advance social good. There is certainly an opportunity to unlock this innovative resource to help nonprofits shape a more equitable, sustainable, and compassionate world. This opportunity calls for a collective commitment to realizing AI’s full potential in finding efficiencies for the greater good – without using AI to replace people.
This opportunity, then, must be undertaken cautiously, with a discerning eye towards the perils that AI presents. To do this, nonprofits should adopt a framework that enhances their capacity to make informed decisions about AI implementation, ensuring its positive impact on the communities they serve.
When thinking about adopting AI, it is vital to cultivate a learning mindset, imagine different possibilities, and establish robust internal policies. It is through this balanced perspective that nonprofits can lead the way in harnessing the potential of AI to drive positive change in the world.
