Will AI soon be reviewing your grant applications? 

As generative AI use spreads, many in the social sector are trying to grapple with its implications. Questions remain about the ethical implications of AI, with many worried about both the quality of its output and its potential effects on the workforce. In a sector committed to advancing the public good, it’s especially important to approach AI adoption with vigilance, weighing its risks and benefits. 

Considering the role AI will likely play in our sector’s future, we at Candid have been looking for opportunities to explore nonprofits’ attitudes toward AI’s growth. In the 2024 Foundation Giving Forecast Survey—our annual survey of foundations—we asked funders whether they had received AI-generated grant applications and whether they would accept them in the future. Curious about whether funders were looking into adopting AI for their own use, in this year’s survey, we asked whether foundations currently use “generative AI to screen applicants or help make decisions about whom to fund” and whether they anticipate using “generative AI to screen applicants or help make decisions about whom to fund in the next few years.” 

Few foundations use AI to screen applications, but some are considering it  

In response to the question about whether they were currently using generative AI to screen grant applicants or help decide whom to fund, 97% said “no,” 1% said “yes,” and 2% didn’t know. When asked whether they expected to use generative AI for those purposes in the near future, however, a much smaller majority of 65% indicated they were not planning to do so. Nearly one-fifth (19%) were considering the possibility of using AI tools to screen applications or help decide whom to fund, another 3% indicated they expected to do so, and the remaining 12% were not certain. This suggests that as many as one-third of respondents have not yet ruled out using generative AI as part of their grant decision-making processes in the near future. 

Comments show multifaceted attitudes and strategies for using generative AI 

We also gave foundations the option of writing in a comment to clarify their responses. We received six comments on the first question and 12 on the second, mostly from organizations that were not currently using generative AI but were considering it.  

We received one comment from a foundation currently utilizing AI: 

• “AI is leveraged to help with grant summaries and generating reporting on applicant trends and our decision trends. AI use will be explored for other ways to help with admin tasks.” 

One participant (the only one that commented and answered “no” on both questions) expressed strong opposition to AI use for screening grant applications or helping make decisions about whom to fund: 

• “Using AI to screen applicants or make grantmaking decisions is deplorable and morally bankrupt.” 

Some of the foundations that are exploring generative AI use are evaluating the risks and the appropriate use cases for AI: 

• “While we see a lot of benefits of using AI, we are approaching internal operational usage in a cautious [manner,] given the significant environmental impact of AI over usage. In terms of leveraging it in our grantmaking processes, while there may be elements we may consider utilizing in process, we would never allow it to make our funding decisions for us without human, expert oversight of the process.” 

Others see generative AI as particularly useful in dealing with more objective parts of the application process: 

• “[Our foundation] is exploring generative AI for scoring the objective parts of applications.” 

One organization expressed that they saw AI as a possible means to expanding their capacity: 

• “We are a small (<$15 million) foundation with only 4 staff—we are often exploring ways of increasing capacity, and if AI can help, we may use it.” 

All in all, the process of generative AI use among foundations is likely to be multifaceted—using AI for some task but not for others—and uneven. While foundations are exploring using AI tools for discrete tasks, there remains a healthy skepticism of whether AI can take over even a minority of grant screening-related work. The written comments suggest “responsibility” and a human element remain important to funders.