Abstract
As generative artificial intelligence (GenAI) becomes more prevalent across all types of organizations, interest in adopting artificial intelligence (AI) tools by charitable nonprofit organizations, who are given 501(c)(3) tax-exempt status by the internal revenue service (IRS), is worthy of attention, given the potential for increased efficiency and raised ethical concerns in resource-drained economies. Guided by the Technology–Organization–Environment (TOE) framework, this study surveys Florida nonprofit organizations (n = 168) to explore if certain factors, such as the external environment, innovative culture, and leadership, influence decisions to adopt GenAI tools in nonprofit organizations. We supplement our survey findings with follow-up qualitative data from 14 nonprofit executive directors. Findings offer guidance to nonprofit leaders for assessing their organizations’ environment and innovative capacities when making decisions on GenAI adoption. Furthermore, there are broad implications for examining the potential of GenAI tools for mission fulfillment and more effective management of nonprofit organizations.
Introduction
Information technology (IT) has been a part of the nonprofit sector for decades (McNutt et al., 2018). As network capabilities advanced in the 1900s, and the commercialization of the internet expanded in the late 1990s, organizations have been early adopters of IT, web2.0/social media, and even artificial intelligence (AI). The nonprofit sector, whose missions center on the common good, is no exception. Although nonprofit organizations are often viewed as lagging in the adoption of new IT, studies have shown that many recognize the strategic importance of digital transformation and have taken steps to integrate generative AI tools (GenAI) into their operational practices (Kanter & Fine, 2022; McNutt et al., 2018; Schöbi, 2024). While early AI was focused on rule-based systems, later advances in machine learning and neural networks have enabled computers to learn just from data. Generative AI is this type of system that can create new content (e.g., text, images, audio, or video) based on inputs (Taeihagh, 2025).
Nonprofit organizations use GenAI for tasks to fulfill their mission, produce social goods, or improve the quality of life in communities. For instance, this may include predictive analytics or machine learning for forecasting donor behavior, natural language processing for chatbots or automated emails, robotic process automation for grant reporting or donor acknowledgements, or recommendation systems for suggesting volunteer opportunities or relevant programs (Kabra & Saharan, 2025; Schöbi, 2024). Some nonprofits use GenAI for data cleaning, digital marketing, and grant management, and as it became more mainstream in 2020, its use has expanded to more advanced functions like analyzing news and social media for advocacy campaigns and even humanitarian aid by considering satellite images to assess disasters (Barzanti et al., 2024; Efthymiou et al., 2023; McNutt et al., 2018). As AI has advanced in the last few years, the possibilities of technology for the sector are significant.
An example of GenAI that is frequently used in nonprofits is ChatGPT, launched by OpenAI, which can generate human-like text and images based on organizations’ specific requests or tasks (Fui-Hoon Nah et al., 2023). OpenAI originally began as a nonprofit organization but is now a commercial organization that offers GenAI tools, such as ChatGPT, that fall under natural language processing and machine learning. It is essentially a chatbot that allows users to ask questions and, in return, provides responses (content generation) to questions based on large datasets (Fui-Hoon Nah et al., 2023). Since it launched, the innovative tool has captured attention across disciplines and sectors, including nonprofit organizations. Amid a dynamic environment and a competitive market, nonprofit leaders must carefully evaluate their organizations’ innovative culture and make decisions regarding technology use to enhance their organizations’ effectiveness and remain appealing to stakeholders. ChatGPT and other GenAI tools can provide nonprofit organizations with avenues for increased engagement, accessibility, communication, accountability, scalability, accessibility, and other benefits (Moro-Visconti et al., 2023; Young et al., 2019; Yuan, Gasco-Hernandez, 2021; Zuiderwijk et al., 2021).
Although there is nascent literature on GenAI use in the nonprofit and public sector literature, fundraisers, grant writers, and social media managers have begun to explore the use of GenAI tools for innovation, organization, and management functions. GenAI ability has been found to effectively overcome various management problems nonprofit organizations face due to human resource limitations (Haynes, 2023). Despite existing potential for nonprofits using GenAI, there are many reasons nonprofits may not use new technologies including costs, human capital and capacities, and mission alignment. Further concerns regarding its use including ethics, privacy, and intellectual property rights may prohibit nonprofit executives from taking advantage of newer technologies. Nonprofit organizations have historically been unable to take full advantage of technology and innovation benefits due to limited resources (Godefroid et al., 2024). In addition, the adoption of new technologies in nonprofit organizations heavily depends on institutional factors, such as supportive leadership and innovative culture (Madan & Ashok, 2023; Zorn et al., 2011).
In this study, we explore whether nonprofit organizations have adopted GenAI tools in their daily operations. Furthermore, guided by the Technology–Organization–Environment (TOE) framework (Madan & Ashok, 2023; Tornatzky & Fleischer, 1990), we examine how certain factors, such as the external environment, innovative culture, and organizational barriers, influence nonprofit leaders’ decisions to adopt GenAI tools in their organizations. To answer these questions, we conducted a survey of 501(c)(3) nonprofit organizations in three large cities in Florida: Orlando, Tampa, and Jacksonville. The survey investigates nonprofit leaders’ knowledge of GenAI tools; their attitudes toward adopting these tools in different management functions, such as fundraising, marketing, and grant writing; and how their organizations’ culture (creativity, competitiveness, or risk-taking) influences their adoption of GenAI. Before beginning the survey, participants were given the following definition of GenAI: “GenAI (generative artificial intelligence) is defined as technology systems that act like humans to solve problems; and GenAI refers to AI models and software that are publicly available and can be modified, improved, or reused by anyone. Some common GenAI are ChatGPT, Bing Chat, and Canva Magic Write.” This means we adopted an inclusive definition of GenAI in our study, recognizing that most nonprofits utilize several GenAI tools, including but not limited to ChatGPT.
Literature Review
GenAI Tools in Nonprofit Management
While AI has been around for decades, it has become more prominent in organizations across all sectors due to the benefits of cloud-based data management platforms, data analysis, and even program evaluation (Azevedo, 2021). The use of GenAI within nonprofits has become more pronounced with advancements in machine learning and data analytics. Nonprofits are beginning to catch up with other sectors in terms of exploring uses of GenAI (Smulowitz & Vogel, 2024), and GenAI is starting to show promise in enhancing operational efficiency and amplifying impact in the past 3 years, since OpenAI launched ChatGPT in 2022 (Androniceanu, 2024).
Several GenAI tools offer nonprofits innovative solutions for data management, donor engagement, and program delivery and increase avenues for fundraising (Alsolbi et al., 2023; Song & Ng, 2023). For instance, machine learning algorithms are utilized to analyze large datasets, allowing organizations to identify patterns and trends that can drive strategic decisions and optimize resource allocation (Holzer, 2024; Song & Ng, 2023). In terms of donor relations, GenAI-powered tools enable personalized communication and targeted fundraising campaigns and can improve donor retention and engagement (Alsolbi et al., 2023). Furthermore, GenAI-driven platforms assist in monitoring and evaluating program outcomes that help demonstrate effectiveness to stakeholders (Cihon et al., 2021; McNutt, 2023). By integrating GenAI into their operations, nonprofit leaders can streamline administrative tasks and improve decision-making processes.
Yet, nonprofit staff are often undertrained and lack capacity on technology support and maintenance at a time when systems and data are vulnerable, and the need for service delivery and effectiveness is increasing (Fine & Kanter, 2020). Research is emerging on ethical concerns of GenAI in philanthropy, specifically over algorithmic bias, openness, and accountability, and ethical rules governing GenAI use have not been developed (Head et al., 2023). Still, nonprofits can play an important role in introducing GenAI into society in new, safe, and ethical ways (Kilicalp et al., 2024). To advance this technological innovation, nonprofit leaders should adopt proactive strategies to assess and incorporate GenAI applications that support their mission-driven goals. Research constantly shows a strong connection between leadership’s understanding and support and the acquisition and utilization of new technologies that support their organization’s mission (Iskandarova & Sloan, 2023; Weill & Aral, 2005). The following sections first review leadership influence on technology adoption in nonprofit sectors and then discuss the key determinants of GenAI use, which can shape leadership’s decision-making.
Leadership Influence on GenAI Adoption
Leadership is a key predictor of an organization’s decision to adopt new technology (Jaskyte, 2011). Hackler and Saxton (2007) emphasize that nonprofit leaders undertake a pivotal role in linking the acquisition and utilization of technologies. Leadership style has been found to contribute to technological innovation. Jaskyte (2011) finds a positive relationship between transformational leadership and technological innovation, suggesting certain types of leaders can drive new programs and activities and facilitate technology innovation (broadly defined in this study as implementing new services, programs, or products in new ways). Other studies have found human-centered leadership approaches like systems leadership are required for implementing AI in nonprofit organizations (Iskandarova & Sloan, 2023).
Organizations with champions of IT and leaders who understand its potential also have greater IT investments (Berlinger & Te’eni, 1999; Hackler & Saxton, 2007; Weill & Aral, 2005). Leadership support of technology is a vital component in organizations that are “IT savvy” (Weill & Aral, 2005). Nonprofit leaders who are supportive of new technology can position their organizations to be more efficient, adaptable, and impactful (Jaskyte, 2011). In studying factors that determine why non-governmental organizations do not use information and communication technology (ICT), leadership engagement and support of IT were one of the primary reasons, along with lack of resources, prioritization, lack of standard software, and concern over implementation risks and data privacy (Godefroid et al., 2024). These findings highlight the importance of having leaders not just on board with innovative practices, but also who are adaptive, collaborative, and informed on new practices and technologies. Still, nonprofit leaders often struggle to manage technology and use technology to adequately position themselves for the future (Ross et al., 2009).
To better understand what factors could influence nonprofit organizations to adopt GenAI tools, we ground our discussion in the TOE framework in the following section.
The Technology–Organization–Environment Framework (TOE)
The TOE framework is an underlying theory to help in understanding advanced technology adoption in nonprofit organizations (Iacovou et al., 1995; Tornatzky & Fleischer, 1990; Yoon, 2024). It is widely used for understanding decisions on technology adoptions in public management (Godefroid et al., 2024; Yoon, 2024). For example, scholars employed this framework to investigate the implementation of E-government (Defitri et al., 2020), blockchain technology (Malik et al., 2021), and Chatbots (Jais et al., 2024) in government agencies. In the nonprofit sector, the framework has been applied to decision-making in several technology integrations, such as the use of digital marketing (Yoon, 2024), communication technologies (Godefroid et al., 2024), and barriers to AI adoption (Kabra & Saharan, 2025).
The framework offers three important components that influence decisions to adopt advanced technology. These components include: (1) T—technology, (2) E—external, and (3) O—organization’s readiness (See Figure 1). We expand the discussion of each component below.
The TOE Framework (Tornatzky & Fleischer, 1990).
The Technology Component
The technology component focuses on the availability and characteristics of the technology to an organization (Cruz-Jesus et al., 2019; Yoon, 2024). External technologies that refer to certain types are available outside the organization but have not yet been adopted due to cost or marketing (Cruz-Jesus et al., 2019; Malik et al., 2021). Internal parts emphasize the technology accessibility in the organization. Both internal and external technologies influence an organization’s adoption decision because they present opportunities and risks that the organization can accept (Amini & Bakri, 2015). The adoption of ICT is a great example. Godefroid et al. (2024) find that technology readiness on the market can shape nonprofit decisions to incorporate new tools in their daily operations. Similarly, when cloud computing systems were affordable and user-friendly, nonprofit organizations, including smaller-sized nonprofits, adopted it quickly to more effectively fulfill their missions. Studies have also shown that when nonprofit members perceive technology as easy to access and user friendly, they are more likely to adopt it (Godefroid et al., 2024; Seo & Vu, 2020). Based on these findings, we propose the following:
The Organizational Component
The organizational component contains several attributes influencing decisions to adopt new technology. For instance, revenue size and staff expertise may influence the adoption of advanced technology (Yoon, 2024). Staff who lack skills, knowledge, and training on technology, in addition to other financial limitations, are less likely to adopt new technology (Defitri et al., 2020). In addition, organizational resources, including revenue (Godefroid et al., 2024; Mao, 2022) and the number of board members (Zorn et al., 2011), can influence the likelihood of advanced technology adoption in nonprofit organizations (Briones et al., 2011; Ihm & Kim, 2021). Lee and Blouin (2019) find that nonprofits’ financial readiness has a positive relationship with their technology management.
Other research suggests that public managers make decisions to adopt new skills or technology within the organizational context. For instance, certain demographic factors may be relevant to technology-related decisions. Most predominantly, financial constraints are a major constraint explaining why nonprofits lag in technology adoption and management (Jaskyte, 2011; Lee & Blouin, 2019; Wallace & Rutherford, 2021). However, other recent studies reveal that certain demographic factors, such as revenue size, personnel numbers, or board size, did not always play a significant role in technology use in nonprofit sectors (Godefroid et al., 2024; Mao, 2022).
Other contextual factors, such as the type of services an organization provides or the subsector it is a part of, can also influence technology adoption for nonprofits. For instance, health and human service organizations may be more cautious in adopting certain technology due to heightened concerns over IT security, data privacy, and compliance with client confidentiality regulations (Barenblat & Cain, 2024). In contrast, nonprofits whose missions rely on public engagement may have fewer incentives to invest in advanced technologies or GenAI than those in engagement-driven fields like culture or education (Xue et al., 2024). Thus, not only does context play a factor, but the nature of the organization (related to its subfield) may influence the way in which new technology is used across nonprofits. Due to this mix of findings, we included the organization’s revenue, board size, and subsector categories as the control variables in our exploration models. The next hypotheses are as follows:
Innovative Culture
Innovation is defined as “the process whereby organizations transform ideas into products, services, or processes” (Baregheh et al, 2009, p. 1333). It is critical in nonprofit organizations because they are facing unique social problems and challenges in a dynamic environment (Dover & Lawrence, 2012). Despite the importance of innovation, the greater challenge is being engaged in continuous innovation (Dover & Lawrence, 2012). Organizations with a strong innovative culture constantly assess, select, and implement new ideas into production (Hughes et al., 2011). This requires the organization to have significant capacity to adapt and put forth resources, which can deploy increased resources and maintain a cycle of doing things in new ways to create value (Fan et al., 2021; Yang, 2012). For instance, an innovative organization has the capacity to tolerate certain risk-taking behaviors, which is important for successful innovation across organization types (Borins, 2001; Brown & Osborne, 2013).
A variety of other organizational factors are related to innovation capacity’s development in nonprofit organizations. For instance, Dover and Lawrence (2012) argue that power is greatly associated with an organization’s continuous innovation. Power from formal authority and relevant expertise influences nonprofit organization’s learning process. Transformational leadership can help to cultivate an influential power in organizational culture to adopt new ideas and respond faster to changes from external environments (Jaskyte, 2004). Therefore, we hypothesize the following:
Communication Process
Communication is critical to nonprofit decision-making processes, and it reflects the organization’s culture and values (Campbell, 2008). The adoption of new ideas, products, or technology often depends on a communication process between employers and employees in an organization (Amini & Bakri, 2015; Suh et al., 2018). Nonprofit organizations often use informal communication to share new ideas, which encourages adoption of innovative products, services, or behaviors (Suh et al., 2018). L. K. Lewis (2014) finds that demands for innovation are usually initiated by key stakeholders through communication to other members. During the adoption process of an innovative idea or service, members in the organization need to announce, persuade, explain, and report information to each other. A strong internal communication can positively influence organizational innovation (Damanpour, 1991; Suh et al., 2018). Thus, we hypothesize the following:
The Environmental Component
The environmental component refers to the external influences and pressures (e.g., competition) that an organization faces (Chan & Chong, 2013; Olfat, 2024). These can be influenced by the external environment, including the current political agenda, public policy, and peer competition. Competition can motivate nonprofit organizations to learn new knowledge, tools, and skills to survive and grow in the market (Weerawarden & Sullivan-Mort). Nonprofit organizations usually face a large mix of stakeholders with different needs and interests in a changing environment. To meet various needs and interests, nonprofits must take proactive steps to deliver services effectively and efficiently (Harrison & Irvin, 2018). Furthermore, nonprofit stakeholders, such as clients and donors, may raise privacy risks and others in using GenAI tools (Willems et al., 2023). Nonprofit organizations may hesitate in their decision of using AI tools based on the stakeholders’ concerns regarding the data privacy issue.
Methodology
Sample
To better understand nonprofit organizations’ opinions and interest in adopting GenAI tools in their operation and management, this study invited 609 registered 501(c)(3) nonprofit organizations from Jacksonville, Orlando, and Tampa in Florida, to participate in an online survey (see the Appendix). We selected these 501(c)(3) organizations with GuideStar Seals (Bronze, Silver, Gold, and Platinum) and contact information listed on GuideStar in 2024. GuideStar is a public charity which collects information about IRS-registered nonprofit organizations. The Seals indicate higher levels of transparency by providing contact information, donation information, mission statements, and leadership information (Candid Inc, 2025). We sent out the online survey invitation to these organizations between July and August 2024 with two waves of email reminders. Among these 609 organization emails, 19 emails bounced back, 3 organizations asked to withdraw, and 168 organizations answered the survey, which yielded a 28.5% response rate.
Variables
Dependent Variable
This study aims to examine whether nonprofit organizations have adopted GenAI tools and better understand their attitudes toward implementation decisions. AI tools are still relatively new in nonprofit organizations’ management and operations. Some organizations indicated that they supported the idea of adopting GenAI in their work, although they were still in the process of finalizing the official decision. To accurately capture intentions of adopting GenAI, we use a categorical dependent variable (5-point Likert-type scale) to ask respondents to what extent they support the decision to adopt GenAI tools in their organizations.
Independent Variables
Following the TOE framework, we organized the independent variables in three sets to explore the research questions. The first set focuses on variables related to organizational readiness to adopt new technologies, the second on technology availability, and the third on the external environment, which may influence nonprofit organizations’ adoption of GenAI tools. In the following section, we present how these independent variables were measured.
Under the organizational component, innovative culture is a key independent variable. To measure innovative capacity, we adopted five measurement items from Fan et al. (2021) and Odoom and Mensah (2019), which measure organizational innovative capacities with a variety of indicators, with minor adjustments for contextual fit (see Table 1 for organizational items and adjusted measurements). Both studies used structural equation models to test validity and reliability and found high composite reliability (CR > 0.7) and average variance extracted (AVE > 0.5), indicating the measures of organization’s innovative culture used are valid and reliable. To ensure internal consistency of this measurement in our study, a factory analysis was also conducted. The KMO value was .65, and Bartlett’s test was significant (χ2(10) = 73.038, p < .001), indicating sampling adequacy. These factors explain 57.83% of the total variance. Communalities ranged from .39 to .98. Reliability analysis showed acceptable internal consistency for each factor (α = .59 to .76)
The Measurement of An Organization’s Innovative Culture.
Because AI tools are still relatively new to nonprofit organizations, the communication process for adopting GenAI tools is at an early stage. Therefore, we employed a dichotomous variable to measure the internal communication process of the decision. We asked respondents if they had communicated with their employees regarding their recommendation on AI tools for work-related tasks (yes/no).
Last, for the organizational component, we included several demographic variables, such as the organization’s revenue size, number of board members, and mission focus, in the model. We collected both revenue size and the number of board members from GuideStar. For the mission focus, we adopted the classification of nonprofit types from the Indiana Nonprofit Project, which utilizes a multi-phase and multi-year evaluation from several databases, including the National Center for Charitable Statistics, GuideStar, and the IRS. The project classified nonprofit organizations into nine major types by their mission focus: (1) Arts & Culture; (2) Education; (3) Environment and Animals; (4) Health; (5) Human Services; (6) International, Foreign Affairs, and National Security; (7) Public and Societal Benefit; (8) Religion and Spiritual Development; and (9) Mutual Benefit (Indiana University, n.d.).
For the Technology Component, we identified three independent variables. The first considers the accessibility of GenAI tools. The survey asked how accessible GenAI tools were using a 5-point scale. The second independent variable identified if the organization has any cost barriers in accessing GenAI tools. We did not ask if the organization allocated a special budget for GenAI tools, because most organizations would record this type of cost in a general category (e.g., technology budget). The last variable under this component asked if the organization lacked technology staff to adopt GenAI tools.
We used two independent variables to observe how environmental components also influence nonprofit organizations’ GenAI adoption decisions. The first independent variable measured if the participating organizations faced external pressure for competition to adopt GenAI tools (5-point Likert-type scale). The second independent variable asked if the organizations had privacy concerns in adopting GenAI tools in their operations.
Analytic Plan
To answer the research questions, the study first used a dichotomous variable to explore the number of organizations that have adopted GenAI tools. We then followed up with several respondents and found that some organizations were strongly inclined to adopt the new technology and were already in the process of doing so, even though they had not yet made any official announcements internally. We also found that the use of GenAI in the respondents’ organizations is correlated to the leadership support of the decision to adopt AI tools (r (168) = .396, p < .01). Therefore, we used leadership support for adopting GenAI (5-point Likert-type scale) as the dependent variable and conducted a Logistic regression. The model is appropriate for analyzing survey data with cross-sectional research designs (Best & Wolf, 2015). Due to the exploratory nature of our work, a stepwise method with a forward selection approach is employed in the model based on the TOE framework. The first model included all factors under the organizational component; the second model added all the factors under the technological component; and the last model included all factors from the environmental component.
To provide a richer context behind the quantitative analysis, we supplemented quantitative findings with qualitative data from the survey’s open-ended question and follow-up conversations with several respondents. We discuss these findings, followed by quantitative data.
Findings
Descriptive Findings
There are 181 organization leaders who participated in our survey. To maintain analytical validity, we only included 168 responses that fully answered the survey. We reported descriptive findings, including the freqeuencies in Table 2 and the central tendency in Table 3. Organizational revenues range from $0 to $402,156,091. Using the typology from the Indiana project (2025), we identified 61 (36.3%) human services organizations, 37 (22.0%) organizations focusing on public societal benefits, 20 (11.9%) arts and culture organizations, 18 (10.7%) health organizations, 15 (8.9%) education focus organizations, and 17(10.1%) other types (including but limited to religion, international affairs, or environmental organizations) in our survey responses. In addition, the average board size of these organizations is 11, with a range of 3 to 40.
The Frequency of the Categorical Variables in the Final Analysis.
The Descriptive Findings of All Variables in the Final Analysis.
Ordinal Logistic Regression Analysis Findings.
Notes: *p<0.1, **p <0.05, ***p<0.01
Findings reveal a growing trend of GenAI adoption among many nonprofit organizations, consistent with other findings on the growth of GenAI adoption across the sector and geographic contexts (Smulowitz & Vogel, 2024). The rise in popularity of GenAI tools leads many nonprofits to pioneer their use across different tasks and functions. In our sample, 60 organizations (35.7%, n = 168) said they have adopted GenAI tools in different operational activities, such as donor engagement, volunteer recruitment, grant writing, and content generation for digital platforms. In addition, 68.7% respondents indicated that they either somewhat support or extremely support the decision of adopting GenAI in their organizations, which indicates a generally positive attitude toward incorporating this new technology. A few leaders (6.0 %) were somewhat against the adoption of GenAI in their organization’s operations, and only three leaders (1.8%) showed that they were extremely against this idea. In the commentary, leaders with opposition wrote: “Writing by AI tools is poorly written, contains outright falsehoods, and no real substance,” and they “Do not trust the reliability of AI and prefer human touch. Everything I have seen so far is that AI advertising is fake-looking. [We are] not ready to use it yet,” and “We are people facing.”
For those 60 organizations, which have adopted GenAI, we also explored how they integrated the tool in their work (see Figure 2). The content generative features are widely adopted by nonprofit organizations for crafting newsletters, social media posts, emails, and grant writing. Some organizations stated that they did not let GenAI tools write the entire email or proposal, but gathered new ideas and information in the writing process. One organization commented that “we do not use the responses of GenAI but instead use the concept(s) offered.” Fifteen organizations in our sample also indicated that they used GenAI tools for program and service delivery, though the extent of risk and exposure to GenAI varied greatly between these organizations. For example, one nonprofit organization stated that they used a legal research platform with GenAI functions to help prepare legal documents for their clients, suggesting that they may be open to greater risk given the nature of sensitive content and processes. Another nonprofit organization uses GenAI tools to develop curriculum for their educational programs. Only seven organizations responded that they use GenAI tools for volunteer management. This included lower-risk activities, such as writing recruitment posts on social media, developing training materials, and giving written recognitions on newsletters or other platforms, among other functions. In the “other” category, several organizations shared that they use GenAI tools for data analysis to help them with strategies for modification or for greater exposure in organizational tasks like decision-making. The data include sales trends, financial information, or certain demographic information (e.g., voters’ behaviors) for organizations specific needs.
What Do Nonprofit Organizations Use AI Tools For?
While many nonprofit leaders showed support for using GenAI tools, not all have communicated this idea in their organization. In our survey results, 47% of the respondents recommended that their staff or board members use GenAI in daily operations or management. We followed up with several organizations and found that the conversations about adopting GenAI were informal. Some nonprofit leaders mentioned that the official decision to incorporate these tools would be on their board meeting agenda in the near future. Others also shared that they have encouraged their staff members to explore the potential of using these tools at work.
The capability of developing an innovative culture can reflect how effectively a nonprofit organization is embracing new technology. Guided by previous work (Fan et al., 2021), we asked survey participants to self-evaluate their organization’s innovative culture using an index ranging from 0 to 25. The findings indicate that all organizations have established a level of innovative culture (the result score ranges from 9 to 25). The average score is 18.72 (SD = 2.83).
Based on the TOE framework, the technological component of this study reflects nonprofit organizations’ accessibility to GenAI tools. Our survey results found that this accessibility was relatively high, with 39.9% of respondents agreeing that generative GenAI tools, like ChatGPT or Bing Chat, were easy to access, and 19.6% strongly agreed. It should be noted that although most organizations indicated that some GenAI tools were easy to use, other barriers to adopting the tools for nonprofit operational purposes were reported. For instance, 20.8% of respondents said the cost of using GenAI tools was a barrier to adoption. Certainly, some specific GenAI tools targeting nonprofit management require financial investment. For example, a 12-month subscription to DonorSearch AI (2024) is over $1,000. 28.6% of respondents also indicated that a lack of technology staff was another barrier to adoption.
Last, in the environmental component, 31% of respondents somewhat felt that they must eventually adopt GenAI tools. This was because they did not want to be behind in the market, and 20.2% of respondents strongly agreed with this statement. We also observed that 28.8% of respondents raised concerns about the privacy issue of using GenAI tools as an environmental barrier to adoption. These nonprofit leaders commented that there were no well-established policies from other organizations or sectors, which they could refer to or use to develop their own policies or guidelines. One respondent wrote that “we have concerns about data security and privacy, and we are struggling with policies and procedures for using GenAI.” Specifically, several nonprofit leaders expressed their concerns about health-related data privacy, given their organization’s context.
Ordinal Logistic Regression Analysis Findings
Guided by the TOE framework, we used ordinal logistic regression and ran three models to better understand nonprofit organizations’ intention of adopting GenAI tools in their operations (see Table 3). Before running the logistic regression models, researchers ensured each observation’s independence. Due to the skewness issue for organization’s revenue size and board members, we used the natural log transformation to model these two variables in the logistic regression. To reduce the multicollinearity issues, we used the Variance Inflation Factor (VIF) to check the models. The VIFs among the independent variables were lower than 5, which is acceptable for the regression models (O’Brien, 2007).
The results of model 1 show that the innovative culture is a significant predictor of the supportiveness of adopting GenAI tools in nonprofits (Estimate = 0.22, Odds Ratio = 1.24, p < .01). This suggests that nonprofit organizations with a lower level of innovative culture are less likely to support the decision of adopting the new tools. Similarly, having a clear communication process can greatly influence the supportiveness of the GenAI adoption’s decision. Those organizations that have not had a conversation about using these innovative tools are less likely to show a supportive attitude (Estimate = -3.20, Odds Ratio = 0.04, p < 0.01). The organizations’ demographic factors, such as the revenue size, board size, and mission focus, did not show statistical significance in the model. Overall, the organizational components in model 1 explained 50% of the variance (Nagelkerke R2 = .50).
In model 2, we added technology components and found communication processes and innovative culture consistently influenced the organization’s intention to adopt GenAI, although these influences were slightly decreased when the technology components were added into the model. In model 2, when organizations’ innovation culture increases one unit, odds of support for GenAI adoption also increase slightly (Estimate = 0.122, Odds Ratio = 1.13, p<0.1). Organizations that do not communicate with GenAI at work showed lower levels of supportiveness (Estimate = -2.70, Odds Ratio = 0.07, p<0.01). In terms of the technology components, accessibility of GenAI tools emerged as a strong predictor of supportiveness toward GenAI adoption (Estimate = 1.12, Odds Ratio = 3.07, p<0.01). Additionally, lacking technology staff is negatively associated with supportiveness (Estimate = -1.14, Odds Ratio = 0.32, p<0.05), which indicates this as an adoption challenge. The cost barrier in technology did not show any significance with the level of supportiveness. Adding technology components increased the explanation of the variance in model 2 (Nagelkerke R² = 0.61).
The last model includes all the components based on the TOE framework, and the overall model performance further improved (Nagelkerke R² = 0.68). This indicates that supportiveness toward GenAI adoption is shaped by many factors. When adding the environmental component, the findings show that peer competition can motivate nonprofits’ support toward adopting GenAI tools (Estimate = 1.32, Odds Ratio = 3.74, p<0.01). Privacy concerns for communities did not show a statistically significant effect. Lacking technology staff (Estimate = -1.02, Odds Ratio = 0.36, p<0.01) and technology accessibility (Estimate = 0.74, Odds Ratio =2.09, p<0.01) remained as a significant influence on the support for using GenAI tools. Similarly, in the organizational component, the communication process also remains significant (Estimate = -2.82, Odds Ratio = 0.06, p<0.01).
Qualitative Findings
At the end of our survey, we asked participants to leave comments on their experiences and opinions on GenAI. Among the 168 respondents, 43 provided open-ended responses. Many nonprofit leaders raised concerns about using GenAI for different tasks. It should be noted that different types of nonprofit leaders have specific concerns. Based on the qualitative data, we classified these findings into four categories, which were mentioned frequently in the survey responses (see Table 5). The classification process follows Lazarsfeld (1955) coding requirements for free responses.
Qualitative Analysis on the Open-Ended Question.
The findings show that 21 organizations (49%) indicated risks of data security and privacy in using GenAI tools; 13 organizations worried that GenAI tools could raise bias and equity issues in their services; 15 organizations (35%) stated concern over whether generative GenAI tools provide accurate and authentic content and information; and 10 organizations (23%) were uncertain about the risks or benefits of using GenAI tools. The increase in concerns over data security and privacy for human service nonprofits makes sense, as these organizations are often collecting information that is personal, identifiable, and legally protected. These organizations are often also serving vulnerable populations, where breaches of information could lead to consequences like loss of housing, legal risks, or personal harm. In addition, human service nonprofits often operate under increased regulations, much like health nonprofits who had the next highest level of concern (43%).
We also followed up directly with respondents who answered the open-ended question, as they indicated a strong interest in the topic and were available to provide additional qualitative responses and clarifications. Fourteen organizations provided additional information about the specific GenAI tools they have adopted in their operations. We summarized that these tools are used for the following: (1) Administrative Support, such as taking notes, modifying emails, or writing assistance. Tools mentioned in this category included: Otter.AI, Grain AI, ChatGPT, Copilot, and Claude. (2) Marking Content Design, such as newsletter design, social media creation, and storytelling video production. The respondents noted various tools, including Canva, MailChimp, and Google Workspace AI. (3) Data Analytic and Donor Management, such as donor data mining, fundraising assistance, and demographic analysis. AI tools like Windfall, Free Will, and Placer AI were highlighted by the respondents.
Discussion
Our findings indicate that the initial idea of adopting GenAI tools can start from internal communication. This is consistent with previous observations on adopting new ideas or technologies, such as cloud-based systems, social media, and mobility from other empirical studies (Harrington & Goodman, 2018; Raman, 2016). Internal communication can help to spread opinions and attitudes about the need for adopting new practices (L. K. Lewis, 2014). Therefore, we suggest that nonprofit organizations should consider including a communication strategy for adopting GenAI, which can state the strategic goals of integrating GenAI tools and detail how they align with the mission. The communication strategy should include dialogue about expectations, experiences, and concerns to help mitigate potential risks.
This study further indicates that innovative culture can be an important factor in shaping decision-making related to the adoption of GenAI tools. Nonprofit organizations with a strong innovative culture better understand the importance of new technologies, which extends beyond new ideas to supportive infrastructure and leadership frameworks and adequate resource allocations (J. M. Lewis et al., 2018). For nonprofit organizations exploring these tools, we recommend conducting a culture assessment to help inform strategic decision-making and ensure mission alignment. This involves reviewing past processes of adopting new technologies and exploring their risk tolerance. Even though public and nonprofit organizations are traditionally viewed as risk-averse (Chen & Bozeman, 2012) and tend to use a “Waterfall” approach 1 to embed new technologies and infrastructure, effective risk-taking and management can lead to successful innovation (Brown & Osborne, 2013). Like many other new tools, technologies, and products, using GenAI tools can bring several risks. Our findings reveal concerns over data security and privacy from nonprofit leaders, and this is evident in practice based on open-ended responses of different levels of risk with GenAI adoption. There were many concerns for nonprofit leaders, and at least some discussed very little guidance from other organizations, networks, and sectors on how to do this in their current operations. Respondents noted that they have not received adequate guidelines on procedures and policies for data protection for their constituents. One respondent stated: “I am concerned with embedded bias and equity in AI tools that are being developed as underrepresented groups are not as engaged in creating these tools, and this can lead to perpetuating inequities. Moreover, so many people seeking to make money off of GenAI tools are involved currently and more needs to be done to safeguard data and information. Some regulations do need to be developed.” These results are also evident in our regression model, which shows that environmental components significantly influence the use of GenAI and lead to a broader discussion on the importance of sensitivity to external political environments in GenAI adoption considerations. Being aware of these risks is necessary; however, taking a wait-to-see approach might stifle an innovative culture. Nonprofits may need to become pioneers in this pursuit to mitigate the unique risks of privacy and security with sensitive data that nonprofits often have for AI use. We encourage both nonprofit practitioners and scholars to continue exploring and developing risk mitigation plans based on their organization’s context. Complementing these efforts, we call for renewed attention to nonprofit governance styles that promote and support responsible risk-taking leadership, ethical organizational culture, and a commitment to implementing appropriate organizational data policies as a mechanism for checks and balances across the sector.
Last, although previous studies have mixed findings on the effects of organizational size on technology adoptions (Mao, 2022), we did not find an association between organizational factors, including revenue, board size, or mission focus, and support for adopting GenAI. Currently, many GenAI tools are still affordable for subscriptive services, making the cost of adoption less of a barrier at this point in time. Nevertheless, as AI technology continues to advance, markets will intensify, which may limit access to adoption. Although we did not find that nonprofit types or subsectors were associated with the support of adoption, our qualitative findings show that different types of nonprofits may raise certain concerns related to their stakeholders’ needs. As the use of GenAI is growing, we expect that some nonprofit organizations with specific mission focuses will build new expectations on the adoption. Both nonprofit practitioners and scholars should keep exploring these new expectations and patterns of the use of GenAI in the field.
Implications for Theory and Practice
Guided by the TOE framework, this study suggests that beyond the organizational component, certain technology factors can also influence nonprofit leaders’ decision of using GenAI in their organizations. Given that some technology is low cost and user-friendly and can be adopted more quickly in public and nonprofit organizations (Godefroid et al., 2024; Slatten, 2012), there is significant potential in practice for leaders to take advantage of tools for different tasks, assuming it aligns with their mission, goals, and risk is managed. Our qualitative finding listed some of these tools, which can meet different needs in various management contexts in nonprofit organizations. Still, leaders from these organizations emphasized the need for capacity building. Lacking skilled staff and expertise was found to be a major barrier in the adoption process. To mitigate this challenge, nonprofit leaders should intentionally develop their technology and GenAI skills and invest in other training and capacity building for their leaders and decision-makers.
Nonprofit organizations are under pressure to reduce administrative costs and compete for funding, potential donors, and other resources (Bunger, 2013). Weerawardena and Sullivan-Mort (2001) find that organizations tend to conceive new ways to deliver greater social impact in a competitive environment. In this study, findings reinforce that peer competition can motivate leaders to explore opportunities for using GenAI and learn new knowledge to improve their organizations’ technology capacity. This competitive advantage allows organizations to focus on developing a sustainable, mission-aligned, and innovative strategy to integrate new tools. Nonprofit organizations may consider viewing others as collaborators, rather than competitors, to share knowledge and experience, manage risks of GenAI, and create greater community impact.
Our work also extends the current literature on the application of TOE by highlighting how the environmental context, specifically peer competition, can act as a catalyst for proactive knowledge acquisition and capacity building within nonprofit leadership. This reinforces the theory that external pressures do not just mandate compliance but can actively drive internal technology readiness and mission-aligned innovation in the sector (Neumann et al., 2024). Still, while this study highlights the importance of the environment, a limitation surrounding the application of the TOE framework is its tendency to treat organizational boundaries as static (Babsek et al., 2025). This can limit how interorganizational collaboration mitigates the effects of competition. Future work could advance the framework by building on how the mission or work of the organization can moderate the relationship between environmental pressure, and also consider longer-term integration of GenAI.
Limitations and Future Work
Our study is not without limitations. The TOE framework provides guidelines to explore potential factors influencing nonprofit organizations’ decisions of using GenAI. The TOE framework itself has some limitations in terms of being descriptive rather than explanatory and has potential for conceptual overlap. For instance, leadership and communication factors may blur between environmental and organizational components, which have been noted in other studies (Godefroid et al., 2024), and we also found this a challenge in our study. Future research may integrate the TOE framework with other complementary perspectives and theories that may address these limitations. Also, because GenAI usage is still new, measurements of the components in the TOE framework may not be comprehensive enough. GenAI is developing at an unprecedented pace compared with other transformative technologies. There might be other factors that play an important role that have not been captured. Future scholars explore additional unique elements or components in the decision-making process of adopting GenAI tools to increase the operational validity and reliability of the TOE framework within these specific organizational contexts. In addition, survey research opens the door to response bias, where participants may inaccurately respond to questions and influence the reliability of the data. Participants may not have fully understood AI and the way the survey was designed inclusively for different types of AI (general, generative, superintelligence, etc.). In addition, employees and volunteers can bring their own devices and consumer technology into the workplace, meaning they may use AI within the organizational context for personal use, though this was not what we aimed to capture in our study. When the study was conducted, GenAI tools were still new to nonprofit practitioners, and there may be a lack of understanding new terminology and its use.
There may also be concern over non-response bias for this population (small, local nonprofits) that may not have the staff size or resources to respond to requests for information over surveys. Furthermore, our survey focused on nonprofit leaders and did not consider perceptions of Board members, staff, volunteers, or other stakeholders that may have different opinions on GenAI adoption. We also recognize that respondents in our sample represent small nonprofits located in populous regions within one state (Florida), and their responses do not reflect the sector at large. Therefore, given generalizability concerns, we would encourage future work to explore larger nonprofit organizations in different regions, as well as other nonprofit stakeholders, and consider empirical work to better understand the complexities of nonprofit operations and adoptions in organizations over time.
Given the critical community-based work that nonprofits are doing in their localities, there is always concern over data privacy in the collection of sensitive stakeholder information. Nonprofits often need to collect information for marketing or service delivery purposes, such as financial data from donors, health information, employment data, legal information, education information, data on minors, and other case-specific sensitive data that require robust data privacy and security measures. For this reason, nonprofits that utilize GenAI must have even stricter security measures to protect data and ensure compliance with data protection laws. They must also be able to articulate decision-making when GenAI is used to ensure transparency and maintain trust. Current U.S. politics may also play a factor in GenAI adoption across all public agencies, particularly in regulation. Thus, even more scholarly attention is needed to understand other external factors (i.e., political and social climate) on potential adoption.
Conclusion
The use of GenAI tools is growing rapidly across organizations and communities, and this study confirms that this trend is also evident in the nonprofit sector. Some nonprofit organizations have taken initial steps to adopt GenAI tools, while others are sitting on the fence, unsure where to fall. To help nonprofit leaders make better decisions, we applied the TOE framework for evaluating GenAI adoption in nonprofit organizations, looking at technology factors, organizational factors, and environmental factors. This approach offers flexibility in understanding adoption for organizations of different sizes and focusing areas and provides a holistic approach to understand various factors influencing GenAI use. Our study also highlights the importance of evaluating the decision of adopting GenAI from multiple dimensions, including the internal and external environment.
GenAI tools are developing fast and becoming prevalent in many aspects of life, including the philanthropic sector. This new technology has brought exciting innovation and opened the door to new possibilities of nonprofit work, but also introduced challenges and confusion across the nonprofit sector in terms of how it can be introduced, employed, and integrated into the organization’s context and external environment, as well as policies guiding its use. This study explored nonprofit organizations’ interests in using GenAI tools to assist with their operational tasks, such as grant writing, marketing, communication, and others. Our empirical findings suggest that nonprofit leaders are taking proactive steps to adopt these tools innovatively, and a growing number of organizations are considering their use. Moving forward, nonprofits will need to balance risk and innovation and may turn to other sectors for guidance on data security and privacy policies for regulating AI use in nonprofits.
Footnotes
Appendix
Acknowledgements
Not applicable.
Authors’ Note
All authors have given consent for the manuscript to be submitted in this current form.
Ethical Considerations
This study was approved by the University of North Florida Institutional Review Board (approval no. IRB-FY2024-74) on May 13, 2024.
Consent to Participate
This study was determined as no human subjects research by the University of North Florida Institutional Review Board. The study conducted an online survey, and a written informed consent was sent to all participants.
Consent for Publication
Not applicable.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
The data of this manuscript are available upon request.