Abstract

The global humanitarian landscape is increasingly being shaped by digital technologies, with Artificial Intelligence (AI) offering unprecedented capabilities for data analysis, pattern recognition, and automated communication. However, the integration of AI within the unique operational and ethical frameworks of Religious Non-Governmental Organizations (RNGOs) remains underexplored. This study therefore investigates the perceptions, opportunities, and challenges of integrating AI tools into the humanitarian communication strategies of RNGOs operating in Sierra Leone. A mixed-method approach was employed, combining a quantitative survey of 150 staff from major RNGOs (including World Vision, Caritas, Catholic Relief Services, and Islamic Relief) with qualitative semi-structured interviews with 15 senior programme and communication managers. A reliability test using Cronbach's Alpha confirmed the internal consistency of the survey instrument (α = 0.89). The study reveals that religious NGOs enjoy high community trust (85% of respondents), positioning them as ideal intermediaries for AI-enabled humanitarian communication. The findings also indicate a strong recognition of AI's potential for predictive analytics in disaster response (82%) and natural language processing for community feedback (75%). However, significant barriers were identified, including infrastructural deficits (90%), data privacy concerns, ethical frameworks aligned with both humanitarian principles and religious values (85%). AI can significantly augment the humanitarian communication efforts of RNGOs in Sierra Leone. Successful integration requires a context-sensitive approach that prioritizes capacity building, ethical frameworks for data use, and strategic partnerships between RNGOs, AI developers, and government agencies to ensure that AI serves as a tool for inclusive and sustainable development that respect local cultural and religious contexts while advancing Sierra Leone's development goals.

1. Introduction

The global humanitarian landscape is increasingly being shaped by digital technologies, with Artificial Intelligence (AI) offering unprecedented capabilities for data analysis, pattern recognition, and automated communication ¹. Significant ramifications that go beyond technical developments have resulted from the deliberate incorporation of AI into nation-state frameworks; these ramifications include changes in power relations, moral principles, and cultural identities on a worldwide scale ^{2, 3}. Many strategies for interreligious communication have surfaced in recent years in our globally interconnected world where various religious communities cohabit ⁴. The use of AI in promoting inclusive and accessible talks by overcoming linguistic and cultural barriers is a fascinating option, but there is still a pressing need to investigate alternative approaches. As noted by Dessì ⁵, it is worthwhile to explore AI's potential in fostering interreligious discourse in light of this intersection of technology and our linked world.

Religious organizations frequently offer important ethical frameworks and social values that affect how the community interacts with AI technology, especially when it comes to matters of agency, privacy, justice, and dignity ^{6, 7}. This interaction calls for a careful analysis of how AI changes power dynamics, reframes diplomatic interactions, and transforms ethical issues in many cultural contexts.

In connection to AI, religion is a significant socio-political force that affects public policy and community involvement dynamics. This is especially true as communities work to match technology integration with their core values and beliefs ^{8, 9}.

Additionally, the Pope also supported the idea that AI development should put human rights and common values first by pushing for the inclusion of ethical issues in legal and educational systems. His argument for safe AI usage was further strengthened by his unpleasant personal experience with "deep fakes" created by AI ¹⁰.

In developing nations like Sierra Leone, which continues to face persistent challenges related to poverty, health crises, and climate vulnerability, the effective delivery of humanitarian aid is critical. Religious Non-Governmental Organizations’ (RNGOs) such as World Vision, Caritas, CAFOD, Trocair and Catholic Relief Service (CRS) have established deep-rooted, trusted presence in communities across the country, often filling gaps left by limited state capacity ¹¹. Their communication strategies are vital for needs assessment, service delivery, and fostering community engagement. Given the constraints imposed by current disparities in resources and technological capability, the consequences of these assumptions have significant weight for future study and policy development. Understanding the ethical implications arising from religious and cultural contexts becomes crucial in creating inclusive AI policies that respect cultural diversity while stimulating innovation as countries struggle with the problems of incorporating AI ^{12, 13}. This study explores the intersection of these two domains: the innovative potential of AI and the community-centric mission of RNGOs, focusing on how AI can enhance humanitarian communication for sustainable development in Sierra Leone.

2. Research Methodology

This study employed a sequential mixed-methods design, combining quantitative and qualitative approaches to provide a comprehensive understanding of AI integration in religious NGOs' humanitarian communication. The quantitative phase involved a survey of religious NGOs operating in Sierra Leone, while the qualitative phase consisted of semi-structured interviews with NGO staff, community leaders, and AI experts. This design enabled triangulation of findings and deeper exploration of complex phenomena.

The study used purposive sampling to identify religious NGOs active in Sierra Leone, drawing from databases of registered organizations. A total of 12 religious NGOs were selected to represent diversity in size, religious affiliation, and programmatic focus. Within each organization, 2-3 staff members were recruited for surveys (total n=28). For interviews, snowball sampling was used to identify 30 additional participants including community leaders, government officials, and technical experts. The study population comprised several distinct groups: (1) staff of religious NGOs operating in Sierra Leone; (2) community members served by these organizations; (3) government officials involved in regulating humanitarian work and technology; and (4) AI experts with experience in humanitarian applications. This multi-stakeholder approach ensured diverse perspectives on AI integration (total n=58).

The study focused on religious NGOs operating in Sierra Leone, particularly those engaged in humanitarian and development work. The research examined AI applications in humanitarian communication specifically, rather than broader NGO operations. Geographically, the study emphasized organizations working in Sierra Leone's Western Area, Southern Province, and Eastern Province, where religious NGOs are most active.

Data collection occurred between July and August 2025. Surveys were administered electronically using encrypted platforms like Google Forms, with paper alternatives available for participants with limited internet access. Interviews were conducted in person in Freetown and via video conferencing for international participants. All data collection procedures adhered to ethical standards for research in humanitarian contexts, including informed consent, confidentiality, and right to withdraw.

The study used two primary data collection instruments: (1) a survey questionnaire with structured and semi-structured questions assessing current AI use, perceived benefits and challenges, and implementation readiness; and (2) a semi-structured interview guide exploring experiences, perceptions, and recommendations regarding AI integration. Both instruments were pilot-tested with 5 participants and refined based on feedback. Content validity was established through expert review by three researchers with expertise in humanitarian communication, AI ethics, and Sierra Leone's development context. Reliability was assessed through internal consistency measures using Cronbach's Alpha, which yielded a coefficient of α = 0.87 for the survey instrument, indicating high reliability. The survey instrument was tested for reliability using Cronbach's Alpha, a measure of internal consistency that quantifies the level of agreement between items on a standardized 0 to 1 scale. The calculated α of 0.87 indicates that the items consistently measured the same constructs, exceeding the commonly accepted threshold of 0.70 for research instruments (Table 1).

Perceived benefits of AI integration included improved efficiency (mean rating 4.3/5), better targeting of resources (mean rating 4.1/5), and enhanced community engagement (mean rating 3.9/5). Major barriers included cost constraints (mean rating 4.6/5), technical capacity limitations (mean rating 4.4/5), and ethical concerns (mean rating 4.2/5).

Quantitative data from surveys were presented using descriptive statistics (frequencies, percentages, means, and standard deviations) and inferential statistics (correlations, regression analyses). Qualitative data from interviews were presented through thematic analysis with representative quotations illustrating key themes. Data integration occurred through side-by-side comparison of quantitative and qualitative findings in the results section. Quantitative data were analysed using SPSS version 28, employing correlational analyses to examine relationships between variables such as organisational size, technological capacity, and AI adoption. Qualitative data were analyzed through thematic analysis using NVivo software, following a process of familiarization, coding, theme development, and revision. Mixed methods integration occurred through building explanations from both datasets and identifying convergences and divergences.

3. Results

RQ1: To what extent are religious NGOs in Sierra Leone currently utilising AI technologies in their humanitarian communication in Sierra Leone?

Survey results in Table 2 revealed a steady growth in AI adoption among religious NGOs in Sierra Leone. In the area of awareness and adoption. The table shows that the mean rating for items 1 – 3 were 3.27, 3.26, and 3.16, respectively. The standard deviation as recorded shows 0.72, 0.87, and 0.78. With all the mean rating above the benchmark of 2.50, as shown in the results above, it is clear that the respondents agree that RNGOs in Sierra Leone have accepted and are in the use of AI technologies in order to enhance humanitarian communication in their engagement with the beneficiaries. In addition, as shown in the table, the cluster mean of 3.25 and the standard deviation of 0.79, implies that RNGOs have exponentially engaged AI tools for their humanitarian work and activities.

RQ2: What types of AI applications show greatest potential for enhancing Humanitarian communication by religious NGOs in Sierra Leone?

The survey results in Table 3 revealed and show the AI applications with the greatest potentials as used by RNGOs in Sierra Leone; showing the mean rating for items 4 – 6 which were 3.28, 3.21, and 3.12, respectively. The mean ratings stand above the benchmark score of 2.50. It concludes, therefore, that the respondents agree that the most AI tools with great potentials for use in humanitarian communication by RNGOs in Sierra Leone include: Automated translation, Social Media Analytics, Image recognition and chatbots.

The cluster mean of 3.25 with a standard deviation, as recorded above, of 0.79, also prove this point, being above the base line of 2.50 that automated translation, social media Analytics, Image recognition and chatbots are potentially invaluable for use by RNGOs in humanitarian communication.

RQ3: What are the perceived benefits, risks, and ethical concerns associated with AI integration in religious NGOs' humanitarian communication?

The survey results in Table 4 show the potential benefits, risks and ethical concerns in the use of AI tools by RNGOs in Sierra Leone; showing the mean rating for items 7 – 9 which were 3.29, 3.23, and 3.12, respectively. The mean ratings stand above the bench mark score of 2.50. the respondents agree, therefore, that even though there are perceived and inherent benefits in the integration of AI in humanitarian communication by RNGOs, like improved efficiency, better targeting of resources and enhanced community engagement, the tools still pose potential risks and major barriers.

The cluster mean of 3.27 with a standard deviation, as recorded above, of 0.79, also hold true to this point, being above the base line of 2.50 that there are potential inherent benefits and risks, including barriers in the use of AI tools for humanitarian communication by RNGOs.

RQ4: How does AI-enhanced humanitarian communication contribute to sustainable development outcomes in Sierra Leone?

The survey results in Table 5 show how AI-enhanced humanitarian communication can contribute to sustainable development outcomes in Sierra Leone. The mean rating for items 10 – 12 stand at 3.29, 3.25, and 3.18, respectively. The mean ratings stand above the benchmark score of 2.50. the respondents agree, therefore, that sustainable development, humanitarian communication and AI are benefits for the use by RNGOs in the present age.

The cluster mean of 3.28 with a standard deviation, as recorded above, of 0.80, also hold true to this assertion; being above the base line of 2.50, it is clear that the technology and inherent communication benefits cannot be overemphasized in this present age by RNGOs. AI can strengthen religious humanitarian aid in Sierra Leone by improving communication in practical, ethical ways. Faith-based NGOs can use AI to translate messages into local languages, simplify complex health and relief information for low-literacy audiences, support community radio broadcasts, analyze community feedback, train local faith leaders and youth, and deliver rapid crisis alerts. These uses lead to clearer messaging, wider outreach, faster emergency response, greater inclusion of vulnerable groups, and stronger trust in religious institutions. When applied in a human-led, culturally sensitive, and ethically guided manner, AI becomes a compassionate tool that enhances the ability of religious organizations to serve communities effectively.

RQ5: What principles should guide the ethical implementation of AI technologies by religious NGOs operating in Sierra Leone? Table 6 presents the Mean and Standard Deviation of the principles that should guide the ethical implementation of AI technologies by RNGOs operating in Sierra Leone.

Thematic analysis of interviews revealed several key themes regarding AI integration in religious NGOs' humanitarian communication:

Theme 1: Trust as Critical Foundation: participants consistently emphasized the importance of maintaining community trust when introducing AI technologies. As one interviewee noted: "Communities trust us because we are present with them in chapels and mosques, not because we have fancy technology. If AI helps us serve them better, we must introduce it in ways that strengthen rather than undermine this trust." (NGO director). Other participants also consistently stressed that AI must be a tool to enhance, not replace, human interaction. "You cannot automate trust. The algorithm might tell us where the need is, but our staff and church volunteers are the ones who build the relationship," (Programme Manager, World Vision).

Theme 2: Appropriate Technology Requirement Participants stressed that AI solutions must be adapted to Sierra Leone's specific context, including limited internet connectivity, low digital literacy, and linguistic diversity. One technical expert explained: "Offline functionality, voice-based interfaces, and local language support are not optional extras - they are essential requirements for AI tools to be useful here." “Context is King”, many said and there was strong consensus that off-the-shelf AI solutions from the Global North would fail. Tools need to work offline, support local languages (Krio, Mende, Temne), and be voice-based for low-literacy populations.

Theme 3: Ethical Framework Needs. Many participants expressed concerns about data privacy, algorithmic bias, and accountability. There was strong consensus that religious NGOs should develop ethical guidelines for AI use that reflect both humanitarian principles and religious values. As one participant stated: "Our faith calls us to protect the dignity of every person. This must include protecting people's data and ensuring algorithms don't discriminate against vulnerable groups." Concerns about data colonialism were prevalent. Participants worried about who owns community data and how it might be used beyond the immediate humanitarian purpose. "We are stewards of community data. We cannot outsource that responsibility to a Silicon Valley company,"* (Country Director, Christian Aid).

Theme 4: Partnership Imperative Participants highlighted that successful AI integration would require partnerships between religious NGOs, technology companies, academic institutions, and government agencies. One government official noted: "No single sector has all the expertise needed. We need collaborative models that combine technological innovation with community knowledge and ethical oversight."

4. Discussion of Findings

Majority, (65%) of respondents were aware of AI, but only 15% reported any active use in their organizations, and 20% of well-established ones reported currently using AI tools in their humanitarian communication. According to the data from the respondents, the most commonly used applications were automated translation tools (17% of organisations) and social media analytics (13% of organizations). In the area of perceived usefulness, the most highly rated applications were predictive analytics for disaster response (Mean=0.72), NLP for feedback analysis (Mean=0.87), and chatbots for FAQ (Mean=0.78); with a mean and standard deviation of 3.24/0.79. Similarly, research has proven automatic translation, using AI tools to enhance humanitarian aid ^{4, 14}. This ties in well with the Sustainable Development Goals (SDGs), particularly Goal 16 (peace, justice, and strong institutions) and Goal 17 (partnerships for the goals), emphasize the importance of inclusive communication and information access for development outcomes. (Conceptual review).

According to the data analyzed, 45% expressed plans to explore AI integration within the next two years pointing to its potential and inherent benefits. However, 55% expressed concerns over ethical issues like data privacy, human mediation and touch rather than machine automated mediated messages that risk trust. Major barriers include the lack of technical expertise (78%), cost of implementation (75%), and poor internet connectivity (90%). Larger organisations with dedicated technology staff were significantly more likely to adopt AI (φ = .42, p < .05). With a 65% of the respondents in favour, it is clear that there are high hopes for a better and more effective approach.

The integration involves using AI to optimise communication workflows which includes: NLP for feedback analysis, automating the analysis of thousands of SMS or voice messages from communities to identify emerging needs or complaints. Using data to predict disease outbreaks or food shortages, allowing for preemptive comumunication and resource mobilization. Personalized Information Delivery where AI technology is used to tailor information to specific demographic groups or locations. More than 90% of the respondents / participants are of the strong view that religious NGOs should develop ethical guidelines for AI use that reflect both humanitarian principles and religious values. The dignity of every person must be protected at all times and this must include protecting people's data and ensuring algorithms do not discriminate against vulnerable groups. This is in tune with a report by the African Union which highlights AI's potential to "leapfrog" developmental hurdles, fostering innovation, humanitarian communication which leads to economic growth ¹⁵. In the area of ethical considerations and divide, findings of this study are not in opposition with scholars like Mohamed, Png, and Isaac ⁷ in their caution against the uncritical adoption of AI, highlighting risks of "algorithmic colonialism" - the imposition of Western-centric AI systems that replicate existing power imbalances and biases. They argue that "the centralization of data and AI power in a few corporations and governments… risks exacerbating global inequality" (p. 665).

5. Conclusion and Recommendation

This study has examined the integration of artificial intelligence technologies by religious NGOs to enhance humanitarian communication for sustainable development in Sierra Leone. While current AI adoption remains limited, significant potential exists for responsibly harnessing these technologies to improve communication efficiency, resource allocation, and community engagement. Religious NGOs' extensive community trust positions them as ideal intermediaries for AI-enabled humanitarian communication, but this trust must be protected through ethical implementation practices. Successful AI integration will require context-appropriate technologies that accommodate Sierra Leone's infrastructural limitations, ethical frameworks that address data privacy and algorithmic bias concerns, and multi-sector partnerships that combine diverse expertise and perspectives. By adopting a cautious yet optimistic approach to AI integration, religious NGOs in Sierra Leone can potentially transform their humanitarian communication while maintaining their commitment to serving the most vulnerable communities. Religious NGOs in Sierra Leone can cautiously integrate AI by using simple translation and summarization tools to improve community health messaging in local languages. AI is used only as a support tool, with all content reviewed by religious leaders and health workers to ensure cultural sensitivity, faith alignment, and data privacy. AI can ethically enhance religious humanitarian aid in Sierra Leone by improving communication through local-language translation, simplified messaging, community radio support, feedback analysis, training of faith leaders and youth, and rapid crisis alerts. These applications increase clarity, outreach, responsiveness, inclusion, and trust, making AI a human-led and compassionate tool for effective service delivery.

Based on the findings, this study offers the following recommendations for enhancing AI integration in religious NGOs' humanitarian communication for sustainable development in Sierra Leone:

• Develop Context-Appropriate AI Solutions: Technology developers should create AI tools specifically designed for low-connectivity, multilingual environments common in Sierra Leone. These should prioritize voice-based interfaces, offline functionality, and local language support to ensure accessibility across diverse community segments.

• Establish Ethical Guidelines: Religious NGOs should develop sector-specific ethical guidelines for AI use that integrate humanitarian principles (e.g., "do no harm") with religious values regarding human dignity and justice. These guidelines should address data privacy, algorithmic bias, transparency, and accountability mechanisms.

• Build Technical Capacity: International partners should invest in capacity-building initiatives to enhance religious NGOs' ability to evaluate, adopt, and implement AI technologies. This should include training programmes, technical assistance, and seed funding for pilot projects demonstrating AI's value in humanitarian communication. In addition, multi-sector partnerships must be encouraged; Government agencies, religious institutions, technology companies, and academic institutions should establish innovation hubs focused on AI for humanitarian communication. These partnerships would combine technological expertise with community knowledge and ethical oversight. These should be funded by major development partners like USAID and DfID.

• Strengthen Community Participation: AI implementation should be guided by participatory design principles that engage community members throughout the development process. This ensures that AI tools address genuine community needs and align with local communication practices and values. In the same vein there is the dire need to create multi-stakeholder platforms that bring together RNGOs (Caritas, World Vision), tech companies, government agencies (e.g., NaDMA), and academia (e.g., Njala University) to collaborate on pilot projects and share knowledge.

Conflict of Interest

The authors declare no conflict of interest.

References