Generative AI: Your Copilot for Grant Writing
Cody Bills, Ph.D.
5/30/2024
Grant writing is a crucial yet time-consuming process. However, generative artificial intelligence (AI), also known as large language models, has dramatically altered the grant writing process. There is some reluctance to adopt AI due to the lack of clear guidelines from funding agencies on its use in grant writing. Academics have the option to dismiss AI—at least for now—or to fully embrace this cutting-edge technology. Let’s consider some benefits and pitfalls of AI, the limited policy that does currently exist from funding agencies, and which AI you should use.
One of the most notable advantages of using generative AI in grant writing is improved efficiency and productivity, allowing for researchers to devote more time to their actual research. Additionally, AI can ensure a consistent writing style and format throughout the grant application, which is particularly helpful for large, complex grant proposals.
Generative AI also excels in analyzing large datasets and incorporating relevant data into the grant proposal. This feature can be especially useful when tailoring grant applications to meet the specific requirements of various funding agencies. By analyzing past successful grant applications, AI can help tailor the proposal to improve the chances of success. Other benefits include summarizing entire research articles and checking for clarity and grammatical correctness, which can be particularly helpful for non-native English writers.
However, the use of AI is not without its concerns. One of the main issues with the use of AI is the lack of a personal touch. Grant proposals benefit from a researcher’s knowledge and perspective to tell the story of the research, something that AI currently struggles to provide. Another potential pitfall is the risk of misinterpretation. Complex instructions could be misunderstood by the AI, resulting in inaccuracies in the proposal.
Generative AI should be viewed as a tool to supplement the grant writing process, rather than a replacement for human input. The quality of the AI’s output is heavily dependent on the quality of the input data. Moreover, it is essential to always review and edit AI-generated content to ensure accuracy and completeness. Citations have proved to be an Achilles heel for generative AI, as there is a real possibility that an AI will incorrectly cite a source or even hallucinate and provide an entirely made-up reference. It is important to be mindful of the ethical implications of using AI in grant writing. Questions of authorship and accountability can arise when AI is used.
Policies regarding the use of AI in grant writing are nebulous. The NIH and NSF have both made it clear that the use of generative AI such as ChatGPT to analyze and review grant proposals by scientific peer reviewers is prohibited. When a reviewer uploads a proposal to an AI chatbot, that proposal becomes publicly accessible, thereby violating the confidentiality requirements of the peer review process. However, there is still no clear guidance on using AI to write grant proposals—at least not until 2025. For now, researchers are encouraged, but not required, to indicate in the project description how and to what extent generative AI was used to develop their proposal.
As mentioned, researchers should be extremely careful about what materials they upload to generative AI. However, the Microsoft 365 license at USU includes access to Microsoft Copilot, an AI powered by ChatGPT 4.0, which does not retain prompts or responses, protecting personal and university data. This makes Copilot an ideal tool for writing projects that you wish to maintain confidential. Copilot is already built into Windows 11, making it easily accessible. Separately, ChatGPT is available as a free version, powered by the older ChatGPT 3.5, or through a paid subscription you can access the significantly more powerful ChatGPT 4.0. The major advantage of ChatGPT is that you can provide instructions and writing examples to create custom versions of AI to assist with specific tasks.
In conclusion, while there are some concerns with generative AI, there are significant benefits to using AI in grant writing. Indeed, about 50% of this article was written using Microsoft Copilot to demonstrate its capabilities. By using AI responsibly and supplementing it with human review, researchers at USU can leverage AI to enhance their grant writing process. This balanced approach can lead to more efficient and effective grant proposals, ultimately benefiting the academic community and the broader society.
To learn more about using generative AI in grant writing, read the article from Seckel, Stephens, and Rodriguez and visit their GitHub repository to view a simple comparison of various AI tools and find helpful prompts for developing more competitive grant proposals. Find more helpful information about using AI on the USU Office of Research website.
The Best-Ever Grant-Writing Training Experience
Bonnie Schenk-Darrington, M.A., MTeC
5/17/2024
Fast forward
How can you kick your proposal writing into high gear? The answer might surprise you!
The best way to learn how to write effective proposals is to become a proposal peer reviewer. This seems counterintuitive. You’re probably thinking that you need to get more experience writing proposals before you can apply to be a reviewer. You are wrong. Having scores of successful proposals is rarely if ever a requirement. In fact, many proposal reviewers are postdocs and early career faculty.
Instead, sponsors want someone who knows their field, communicates well, and is willing to follow instructions. Besides technical knowledge, sponsors may ask for experience in tasks such as teaching, developing curriculum, program management, or fiscal management/budgeting. Does the science make sense? Is the program doable? Is the budget reasonable? Do the applicants have the equipment, facilities, knowledge, and experience needed to do this project?
There are many great reasons to become a proposal reviewer. Besides giving valuable service to a sponsor, you make connections with like-minded colleagues and program officers. As you review proposals written by others, you gain a more practiced eye for what works and what doesn’t—and that benefits your own proposal writing skills. Especially if you’re reviewing proposals for a program you hope to apply for, it becomes on-the-job training for your own future proposal efforts.
Generally, you must submit some kind of application to be considered. Sometimes, there will be a form to fill out and sometimes you’ll just email a resume or biosketch to a program officer. Because panels want diverse applicants, you may be asked questions about your ethnicity and race. If you are chosen, you will generally receive training or orientation and have a chance to ask questions before you begin reviewing proposals.
You may or may not receive a stipend. If travel is required, the sponsor will generally cover it. Sponsors generally appreciate reviewers and try not to overburden any one reviewer. The firsthand knowledge you gain will pay dividends.
Here are some links to sponsors’ reviewer applications and process explanations.
-
USDA NIFA (You can also find a great article about being a grant reviewer for various types of USDA opportunities here.)
Some other government agencies, such as the Utah Dept. of Agriculture and Food, the Utah Office for Victims of Crime, NOAA, and the EDA post online about their need for reviewers on a program-by-program basis (click on each agency name for a sample request for reviewers from the past). If there’s a particular sponsor/program you’d like to review for, the best thing to do is search them online and just ask the program officer if they’re accepting applications for reviewers. You should also subscribe to their newsletters where applicable so that you know when new grant opportunities (and probably review opportunities) come available.
Wherever you’re at in your career, becoming a proposal reviewer might be just the experience you need to give your grant-writing skills a boost.