Insider Brief
- A Clemson University study found AI-generated nonprofit fundraising appeals performed about as well as human-written appeals in emotional persuasion, donation intent and perceived message quality.
- AI-generated guilt appeals produced stronger feelings of guilt than human-written versions, while empathy-focused appeals from AI and humans performed similarly.
- Researchers said the findings suggest AI could become a low-cost communication tool for nonprofits with limited budgets and staffing resources.
Artificial intelligence may now be capable of writing nonprofit fundraising appeals that persuade audiences nearly as well as trained human communicators, according to a new study that tested whether readers reacted differently to emotional messages written by people or AI systems.
The study, published in the Journal of Nonprofit & Public Sector Marketing, found that AI-generated appeals matched human-written messages across nearly every major category the researchers tested, including empathy, donation intent and perceptions of ad quality. In just one exception, AI-generated guilt appeals actually produced stronger feelings of guilt among readers than human-written versions.
The findings arrive as nonprofits increasingly experiment with generative AI tools such as chatbots and large language models to reduce marketing costs and produce fundraising materials more quickly. The researchers said the results suggest AI may offer smaller and underfunded organizations a practical way to create persuasive communications without relying heavily on outside agencies or consultants.
The study was conducted by researchers Carlina DiRusso and Virginia Harrison, both of Clemson University, who examined whether AI systems could effectively create emotional appeals tied to charitable giving. Their broader question centered on whether AI can successfully imitate the emotional intelligence typically associated with human communicators.
To test that idea, the researchers designed an online experiment involving 498 U.S. participants recruited through the Prolific survey platform. Participants were shown one of several nonprofit-style appeals related to hunger in the United States and then asked about their emotional reactions, donation intentions and perceptions of message quality.
The experiment compared messages written by three human communication professionals against appeals generated by three AI systems: Claude, Copilot and Gemini. Each system and each human writer produced both a guilt-focused appeal and an empathy-focused appeal centered on food insecurity and donations to Feeding America.
AI Source Bias
Participants were never told whether the messages were written by AI or humans. The researchers said this was a deliberate attempt to remove “source bias,” a growing issue in AI communication research where audiences react differently once they learn content was generated by artificial intelligence rather than a person.
Previous studies often found humans outperforming AI in emotional communication, but many of those experiments disclosed the content creator beforehand. The Clemson researchers argued that real-world audiences frequently do not know whether nonprofit emails, advertisements or donation letters were created by AI systems.
The study found that empathy-focused appeals written by AI and humans produced nearly identical emotional responses. AI-generated empathy appeals scored slightly higher on donation intention measures than human empathy appeals, though the researchers cautioned that result should be treated as exploratory rather than definitive because the broader statistical test was not significant.
AI-generated guilt appeals produced significantly stronger feelings of guilt than human-written guilt appeals. Researchers suggested AI systems may have been better at structuring language around themes of moral obligation and responsibility, which are central to guilt-based persuasion models.
The study leaned into the “Emotions-as-Frames Model,” a communication theory proposing that emotional reactions shape how audiences process information and make decisions. In nonprofit marketing, guilt appeals often attempt to make audiences feel they have ignored a moral obligation, while empathy appeals encourage readers to imagine the experiences of people in need.
Empathy and Emotions
Researchers found empathy was generally the more effective emotional strategy overall.
Empathy-driven appeals increased both donation intentions and issue involvement, meaning readers became more interested in the underlying issue itself. Guilt, meanwhile, increased donation intentions but did not significantly raise issue involvement.
This factor could matter for nonprofit organizations trying to balance immediate fundraising goals against long-term public engagement campaigns.
The researchers said empathy appeals may help organizations build broader commitment to causes because they encourage perspective-taking and emotional connection. Guilt appeals may still be useful for urgent fundraising campaigns, but they can also create negative reactions or “reactance” if audiences feel manipulated.
The study also examined several measures commonly used in advertising research, including perceived message quality, ad liking and perceived message sensation value, a metric tied to stimulation and emotional intensity.
Across those categories, AI and human-written messages performed similarly. Participants generally rated empathy appeals more favorably than guilt appeals regardless of who or what produced them.
How AI Can Help Nonprofits
The researchers reported the findings could have implications for nonprofits facing budget constraints and staffing shortages.
According to the paper, nonprofit organizations operate with limited resources and increasingly need to prioritize mission-related work over time-consuming communication tasks. AI-generated writing tools, the researchers suggested, could help organizations automate portions of fundraising and donor outreach while allowing staff to focus on strategy, relationships and program delivery.
The researchers stopped short of advocating for fully automated nonprofit communication systems. Instead, they framed AI as a collaborative tool that could assist human communicators rather than replace them entirely.
The study also pointed toward broader changes in digital fundraising and donor engagement. As nonprofits rely more heavily on online campaigns, social media and virtual donor interactions, AI tools may become embedded not just in written appeals but also in chatbot conversations, personalized solicitation strategies and automated marketing systems.
Limitations and Future Research Directions
The researchers reported that, like most studies, there are limitations to consider and areas for future work.
First, the participant pool skewed relatively well educated and predominantly White, which could affect how broadly the findings apply to other populations. The researchers also noted that issue involvement may be difficult to change through exposure to a single message, potentially limiting the observed effects.
Another limitation involved audience awareness. Although participants were not informed about whether messages came from AI or humans, the study did not ask whether respondents suspected AI involvement. Future studies should test whether readers can detect AI-generated emotional writing and whether those perceptions alter persuasion outcomes, the researchers said.
The paper also raised ethical questions surrounding AI-generated nonprofit messaging, particularly in emotionally sensitive contexts involving poverty, hunger or crisis response. Researchers said future work should examine how organizations make decisions about using AI in highly personalized fundraising situations and whether donors react differently once AI involvement becomes more visible.
Despite those caveats, the study adds to a growing body of evidence suggesting generative AI is rapidly improving at tasks once viewed as uniquely human, including emotionally persuasive communication. For nonprofits struggling with shrinking budgets and rising competition for donor attention, that could represent both an opportunity and a disruption.
