AI Can Write the Grant. It Can't Build the Organization.
Systems scientist Dr. Dorian Wilkerson challenges the AI grant-writing boom, arguing that persuasive proposals cannot replace the infrastructure, capacity, and measurable readiness required for sustainable funding.
ATLANTA, Nov. 18, 2026 /PRNewswire/ -- Wilkerson Publishing Group today highlights the work of systems scientist, statistician and author Dr. Dorian Wilkerson, who challenges the growing AI grant-writing boom, arguing that a persuasive application does not prove an organization has the infrastructure or capacity to deliver what it promises.
Recent research illustrates the distinction.
Elsevier's 2025 Researcher of the Future survey of 3,234 respondents across 113 countries found that 41% used AI tools to draft grant proposals.
An August 2026 study in the Proceedings of the National Academy of Sciences analyzed approximately 5,700 confidential National Science Foundation and National Institutes of Health proposals from two U.S. research universities. Higher estimated AI involvement was associated with an approximately four-percentage-point increase in NIH funding probability, but no statistically significant association emerged at NSF. These observational findings do not establish causation.
Funding success and project feasibility are different questions. An award reflects a funder's investment in a proposed plan; execution depends on financial resources, staffing, operational capacity and measurable systems.
A 2025 pilot study in Innovative Surgical Trends tested one ChatGPT-generated surgical research grant application, including hypothetical data. Two blinded reviewers considered it well written, but one rated it unacceptable because of feasibility concerns. This single-proposal experiment cannot establish general failure rates or measure nonprofit readiness.
The grant studies examined research proposals, not community nonprofit programs. Wilkerson applies their underlying questions to broader funding architecture: Can an organization document need, test its approach, staff delivery and measure results? AI can support these decisions, but leaders remain accountable for implementation.
Consider a hypothetical program promising weekly 30-minute coaching sessions for 400 families. That requires 200 direct-service hours each week. A staffing plan providing only 60 hours cannot meet the commitment, regardless of how convincingly the proposal is written.
Wilkerson describes the difference between promised capacity and demonstrated capacity as a conceptual "fundability gap."
"AI can write the promise. Only the organization can keep it," Wilkerson said.
His book, Grant Stacking: Building Funding Infrastructure That Scales, introduces four stages: Infrastructure, Capacity, Fundability and Funding. The framework encourages organizations to examine readiness before pursuing additional resources.
"Funding follows infrastructure," Wilkerson said.
Organizations can explore the approach through the free, 16-question Grant Stacking Readiness Lab at https://dorianwilkerson.app. The self-reported assessment examines five dimensions of readiness and does not predict grant awards.
About Dr. Dorian Wilkerson
Dr. Dorian Wilkerson is a systems scientist, statistician, author and founder of Mordenwood National Network, LLC. He holds a Ph.D. in Systems Science and Engineering. His work connects statistics, systems engineering, organizational infrastructure and funding strategy.
Media Contact
Dr. Dorian Wilkerson
Wilkerson Publishing Group
(770) 647-9824
[email protected]
https://dorianwilkerson.org
SOURCE Wilkerson Publishing Group


Share this article