AI-generated applications and papers are overwhelming academic funding and publishing systems. Grant submissions have surged across 12 major European funders: Marie Curie fellowships received 22,000 proposals in 2026, three times the 2022 total, while ERC starting-grant applications rose 60%. The growth tracks increased use of AI tools that can produce proposals at extraordinary speed. Journals face a similar flood, including 20,000 new titles in a year, sharply higher submission volumes, suspiciously rapid reviews and revisions, and papers attributed to nonexistent authors or real researchers who never agreed to participate. Editors and funders are turning to AI screening, but filtering poor submissions cannot solve a problem whose volume keeps expanding.
The collapse of conventional publishing is welcome insofar as it threatens a corporate system built on unpaid academic labor and publicly funded subscriptions. But replacing proposal-driven grants with funders’ priorities and AI-based selection raises deeper concerns: bureaucrats may determine research agendas, AI-generated work could make it impossible to distinguish researchers’ abilities, and human academics may appear unnecessary altogether. The argument challenges the assumption that AI’s spread is inevitable and that institutions must simply adapt. If AI agents can write grants and papers, they may eventually create and evaluate research themselves, leaving universities and researchers vulnerable to political and commercial choices. The resulting disruption may end a dysfunctional system, but it offers no assurance that a fairer or more intellectually diverse one will follow.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.