Even as Artificial Intelligence (AI) surges in popularity, with 73 percent of internet users in India exposed to it in some capacity, the social sector remains at a nascent stage of AI adoption. An informed and intentional approach to building AI can potentially help nonprofits leverage data-driven decision-making (such as analysing datasets and identifying trends) or enhance efficiency through process automation. However, any move towards full-scale adoption is first confronted with two important questions: How are the AI needs of nonprofits different, and what do these organisations think about AI? Only after tackling these questions can other crucial ones—such as how nonprofits can develop AI to suit their needs, and what an ethical and just adoption looks like—be answered.
To address some of these queries and investigate the deployment and usage of AI in the nonprofit landscape, GivingTuesday conducted an AI readiness assessment. Based on a survey of 251 organisations across six regions, it highlights how nonprofits in India are utilising AI, their comfort levels with various AI tools, and the challenges they experience. The India sample was a subset of a larger global study, which helped conduct a worldwide analysis.
The study included small, medium, and large organisations from rural, urban, peri-urban, and metro areas. Most of the organisations worked on the ground with local communities. Just 4 percent of respondents from these organisations were technical staff or involved in monitoring, research, evaluation, and learning (MERL), whereas more than half of the sample comprised senior leaders and co-founders.
Among the surveyed organisations, 55 percent had used generative AI to write text or create images. In contrast, 30 percent of organisations had never used any form of AI. The breakdown and key findings of the report can be found here.
The survey’s main findings are as follows.
1. AI readiness of organisations is weakly dependent on capacity
The study sought to evaluate if an organisation’s capacity could predict its likelihood of adopting AI. To do this, it assessed organisational capacity and AI readiness using the following factors:
- Capacity was evaluated considering an organisation’s staff size, age, geographic reach, and area of expertise. These factors reflect whether the organisation can absorb funding and implement small to large interventions.
- Readiness was calculated based on the data collected by the organisation, presence of technical staff, data policies, and current utilisation of AI.
Based on this data, the study found that high-capacity organisations are better positioned to improve their AI readiness and would benefit more from such improvements. But overall, organisational capacity is not a strong predictor of readiness.
The study also examined if having a Foreign Contribution (Regulation) Act (FCRA) license had any correlation with an organisation’s AI adoption. More than half of the organisations in the sample had FCRA, but no clear relationship could be established between having the license and an organisation’s capacity or AI readiness. However, the survey found that nonprofits with FCRA were collecting data in more varied forms and were slightly more likely to be using some form of AI.
2. Early vs late adoption determines AI usage behaviour
The surveyed organisations were divided into two categories: early AI adopters and late AI adopters. Early adopters tended to be larger organisations with greater resources and infrastructure. These were predominantly located in urban areas, had more staff, and demonstrated higher levels of data collection and technical capacity. These factors enabled them to experiment with and implement AI in more advanced ways. Such organisations were also more likely to have formal data-use policies, cloud-based data storage, and technical personnel such as IT staff or MERL experts—all of which contribute to their ability to adopt AI more effectively.
Only 10 percent of late adopters expressed interest in using AI for applications beyond generative AI.
In contrast, late adopters of AI were smaller organisations with fewer staff members and constrained budgets. They were based in rural or peri-urban areas, where access to digital infrastructure is limited. Such organisations were less likely to have data-use policies or cloud storage, and they often didn’t employ dedicated technical personnel. As a result, their engagement with AI was found to be limited, with most using AI in simpler, more accessible ways, if at all. It isn’t surprising, then, that generative AI, which requires less specialised knowledge, was often their first experience with AI technologies.
However, despite the early adopters being more tech-savvy, their current data use remains manual and tabular, which suggests that they have a long way to go when it comes to AI adoption.
Only 10 percent of late adopters expressed interest in using AI in the future for applications beyond generative AI, and 40 percent did not know what they would like to use AI for. As a result, most other AI use cases lagged behind, with only slight uptake observed for chatbots and transcription.
Early adopters were more experimental—approximately 60–80 percent wanted to use AI for features such as virtual assistants, data interpretation, prediction, chatbots, transcription, and more. In fact, most had experimented with AI in three or more of these ways. They were also found to have a greater appetite for expanding AI usage, with intentions to scale it two or three times more than late adopters.
Overall, early adopters see AI as offering more rewards than risk and are more comfortable using it for work. Late adopters feel differently, with many stating that they either don’t know how to evaluate the risks or perceive the risks and reward as equal. Around a quarter of both groups felt neutral about their comfort level with using AI.
3. There are some apprehensions about AI
Indian organisations hold a mix of optimism and apprehension regarding AI. Nonprofits were hopeful that AI can bring efficiency and productivity to their work, but they also fear that it could lead to dependency, skill gaps, and job displacement.
Organisations had different hopes and fears based on their area of expertise.
In addition, respondents were optimistic about the role of AI in facilitating data analysis and decision-making within the organisation. However, they were concerned about the lack of knowledge and training with respect to AI, as well as issues around data privacy, security, and bias and fairness. A few nonprofits were also worried that AI might impact human creativity and originality. Further, they mentioned apprehensions about the ethical implications of leveraging AI in everyday operations and its broader societal implications.
It is interesting to note that organisations had different hopes and fears based on their area of expertise. Those working in education expressed confidence that AI could enhance opportunities through personalised teaching and by automating administrative tasks that overburden staff. However, they also expressed concerns that students’ over-reliance on AI could reduce creativity and critical thinking.
Nonprofits working in community development saw AI as an opportunity to optimise resource allocation and identify needs. But they were aware that AI could perpetuate existing regional inequities. Organisations working towards women empowerment echoed the fear of AI worsening gender biases. Health organisations thought that AI could improve diagnostics and treatment for marginalised groups but were worried about the privacy of sensitive health data.
Indian organisations’ take on AI: A global comparison
- Indian organisations were twice as likely than the global sample to have technology or data staff. Globally, hiring a MERL and/or tech person was associated with a higher likelihood of using AI, but this was not the case with India. In fact, Indian organisations with a tech person didn’t score highly on AI readiness.
- Compared to the global average, Indian organisations were more comfortable using AI at work—29 percent gave it a score of 10 on a scale of 0–10.
- Indian organisations focused more on the benefits of AI and had fewer concerns about data protection/privacy issues. Therefore, they are more likely to share data without having data-use or sharing agreements in place compared to the global average. The research suggests that this might be because in India, the culture and legislative landscape around data protection regulation is at a nascent stage.
Given these findings, the study suggests that nonprofits should first explore AI’s relevance and potential use cases for their own organisation. It is important to provide targeted support to smaller, resource-constrained nonprofits that wish to adopt AI. Enhancing their digital infrastructure, improving access to technical expertise, and developing capacity-building programmes focused on AI adoption could help bridge the gap and create meaningful pathways for all types of organisations to fully leverage the potential of AI. However, at the same time, it is necessary to establish safeguards as the vast majority of those using AI are using technology products managed by others. Understanding the nuances in current AI adoption and knowledge is integral to achieving equitable and beneficial AI adoption for the social sector.
—
Know more
- Read the full report by GivingTuesday.
- Learn more about how the education sector in India can adopt AI.
- Learn more about nonprofits that are already using AI.