A 62% surge in applications to the European Research Council’s prestigious Starting Grant programme could push its success rate below 5%, increasing concerns that strong research proposals will go unfunded. Researchers say the rise is partly linked to the growing ease of producing grant applications with artificial-intelligence tools, which is placing additional pressure on reviewers and selection panels.
When the success rate of a grant programme falls below a certain threshold, it essentially becomes a lottery.Credit: Javier Rodriguez Jimenez/Shutterstock
“These numbers are a bit frightening,” said Maria Leptin, president of the European Research Council (ERC) in Brussels. She said it was becoming increasingly difficult for staff and reviewers to handle such a large volume of applications.
The ERC’s Starting Grant programme provides early-career scientists with base funding of up to €1.5 million (US$1.7 million) over five years. By the 14 October deadline, it had received 7,710 proposals for the 2027 grants, compared with 4,807 the previous year.
A maximum of 373 grants is available, down from 421 the previous year because of funding constraints. The success rate for the 2027 grants is estimated at less than 5%, or about 4.8%, compared with 8.8% previously. “It’s almost a lottery,” said Devan Mehta, a systems and synthetic biologist at the Catholic University of Leuven in Belgium who applied for a Starting Grant this year.
Source: European Research CouncilSource: European Research Council
Applications surge
The ERC is the European Union’s leading funder of basic research, with an annual budget of almost €3 billion. It offers three grant categories for individual researchers at different career stages: Starting, Consolidator and Advanced. Applications to more than one of these categories have increased significantly in recent years.
Lynn Kamerlin, a molecular biologist at the Georgia Institute of Technology in Atlanta, says similar increases in research-grant applications are occurring worldwide. Research has examined grant success rates.
Kamerlin says the ideal success rate for a grant programme is about 20%. Below that level, it becomes increasingly difficult to identify the strongest research without relying on subjective judgement, which can leave a substantial amount of good science underfunded. “People don’t know the difference between the top six, eight and 12 percent,” she said.
One possible approach, she suggests, would be for a panel to select the top 20% of applications and then choose the winners randomly, as some funders do. Anything below that standard, she says, “is just a lottery ticket”.
AI is contributing to the increase in applications by making grant proposals easier to write. “Text generation has become very easy,” Kamerlin said. She said this was placing tremendous pressure on the system because AI-generated applications can contain claims that are “complete nonsense”, requiring reviewers to scrutinize proposals more closely.
Source: www.nature.com