Researchers and higher education leaders warned that newly emerging AI agents could pose an “existential threat” to how grants are awarded, arguing that assessment systems have not kept pace with rapid technological change. The concern centers on whether grant and evaluation workflows can reliably interpret outputs generated or influenced by agentic AI tools. The warning points to a mismatch between AI development and policy or review methods used in research funding decisions, particularly as evaluators struggle to determine authorship, originality, and compliance. For universities competing for federal and philanthropic funding, the alert signals that proposal review practices—and the evidence requirements for supporting claims—may need sharper guardrails.