The U.S. Department of Education has a policy explaining how grant applicants can use artificial intelligence to create proposals. Some universities are issuing new grant-specific AI tools, along with courses and workshops on how to best use AI in grant writing. Suffice to say education professionals have already started incorporating AI into the grant application process, but some experts caution against trusting those tools to do too much.
Jeanne Law, an English professor at Kennesaw State University in Georgia who teaches an online course about AI for grant writing, said the way applicants use AI makes a big difference. She said education professionals sometimes equate the grant application process with writing, but it involves a lot more than that. Many federal grants recommend at least three months of preparation time, and AI can expedite some of the process.
IDENTIFYING FUNDING OPPORTUNITIES
The first step is identifying suitable funding sources. One company stepping up to help people with that is Granter.ai, an AI-powered platform that helps companies find and apply for grants across Europe, and, soon, the United States. CEO and co-founder Bernardo Seixas said that the grant landscape in the U.S. is less structured than in Europe, with a variety of private, nonprofit, federal, city and state grants. Applicants rely on a mix of government databases, like grants.gov, foundation websites, subscription-based funding directories and professional networks to find opportunities that match their needs. This requires applicants to manually sift through listings, interpret eligibility criteria and keep track of deadlines. AI tools such as Granter.ai aggregate these sources and expedite the matching process.
“Even if you know they exist, which is a big if, you still need to read through all the legal jargon of these grants to make sure that you are eligible to claim them,” Seixas said. “We are also fixing that by having AI read all the eligibility criteria.”
The portal then matches qualifying applicants with funding opportunities automatically.
GRANT WRITING
Grant applications typically require written pieces, including a problem or need statement, as well as a budget, specific written objectives and a plan to accomplish them. The problem statement is persuasive writing, trying to impress upon the funding organization that the issue is important, and the applicant can achieve what they want to do with the funding.
Some AI tools can take in the requirements of the grant and ensure the application includes key pieces of information. And some tools, such as Granter.ai, are trained specifically on a database of grant applications, unlike general large language models (LLM).
For experienced grant writers, Brian Quinn, a librarian at Texas Tech University, said a free tool like ChatGPT may be enough to assist with the writing process. However, for those newer to grant writing, AI tools that are designed for the purpose might be a better fit, requiring less knowledge of prompt engineering.
Law said AI may be less helpful with the actual writing process than with other steps of grant applications. One reason is the need for a human to think through and explain the project, and the concept of “garbage in, garbage out”: If the project is not well thought out, the narrative is not going to be convincing, no matter what tool the applicant uses to help.
Another is that large language models are designed to please the user. If asked to write something persuasive, an LLM might come on too strong unless the user finds a way to implement a guardrail through prompts.
Law said it is easy for her to spot AI-generated grants in her classes. Though she calls herself an “AI evangelist,” she is not convinced AI will improve much in this arena, even when someone eventually creates artificial general intelligence, or AI’s ability to reason at the same level as a human.
“You can reason all day long, but persuasion and the ability to use not just logic but emotion and credibility to persuade someone, especially persuade someone to fund a project, to give you money — that, to me, is a uniquely human capability that I don’t think even artificial general intelligence is going to be able to match,” she said.
FOLLOW UP
After grants are awarded, recipients are often required to give updates on expenses and progress on project goals, Seixas said. AI tools such as Granter.ai can help keep track of what kind of follow-up is required using information the tool has tracked since the outset and aggregate data for these reports. Applicants can also use these insights in future application cycles.
“AI is a great partner,” Law said. “It’s a great collaborator, but you always have to be the human at the helm.”
