SoylentNews

upstart [soylentnews.org] writes:

████ # This file was generated bot-o-matically! Edit at your own risk. ████

Fearful of AI-generated grant proposals, NIH limits scientists to six applications per year [science.org]:

Scientists hoping to obtain some of the National Institutes of Health’s (NIH’s) dwindling research funds face a new challenge: They will be limited to submitting six applications per calendar year, according to a notice [nih.gov] the agency released this week. The policy, which also prohibits applications written with the assistance of generative artificial intelligence, is ostensibly designed to prevent researchers from overwhelming the NIH grant-review system with large numbers of proposals, especially low-quality ones produced with AI.

But some critics worry the cap—which applies to grant resubmissions, renewals, and revisions as well as original applications—will hurt scientists who are already struggling to obtain federal funding as NIH freezes and rescinds many grants for political reasons and President Donald Trump’s administration seeks to cut the agency’s annual budget by more than one-third [science.org].

Several researchers have taken to social media to express concerns about the new limit. “Many of us are going to be screwed over by this new policy [bsky.app],” Brian Stevenson, a microbiologist at the University of Kentucky, lamented in a post on Bluesky.

Others, however, argue the cap is warranted—and perhaps even necessary. “It’s a reasonable approach to an unfortunate problem,” says Michael Lauer, who served as deputy director for extramural research at NIH until his retirement [science.org] in February. (NIH did not provide a current official for a requested interview on the new policy.)

Lauer notes that not long before he left NIH, he and his colleagues identified a principal investigator (PI) who had submitted more than 40 distinct applications in a single submission round, most of which appeared to be partially or entirely AI generated. The incident was “stunning” and “disappointing,” says Lauer, who was not involved in creating the new NIH policy but hopes the cap will discourage other researchers from abusing the system.

Aside from the cap, the policy makes clear that NIH will not consider AI-generated proposals to be the original work of applicants. “NIH will continue to employ the latest technology in detection of AI-generated content to identify AI generated applications,” the agency’s notice says. If AI use is detected after an award has been granted, NIH warns, the agency may refer the matter to the Office of Research Integrity while imposing penalties. It’s unclear whether applicants will have an opportunity to appeal these decisions, or which tools will be used to detect AI-generated content. These programs can vary wildly in terms of accuracy, with some showing bias against non-native English speakers [theguardian.com].

Attempts by federal science agencies to cap grant submissions and awards are not without precedent, although these policies have historically been unpopular. In 2017, for example, NIH introduced—and then quickly abandoned [science.org]—a plan to limit the amount of support any individual scientist could receive. Researchers feared the cap would harm collaborations and large, productive labs. And in 2018, the National Science Foundation introduced a new policy limiting biologists to only one proposal per year, generating fierce pushback from the research community [science.org]; the agency lifted the restriction [science.org] the following year.

The tricky part, Lauer says, is determining an appropriate cap. Although he argues that a single researcher submitting more than 40 applications is “clear abuse” that wastes valuable time among NIH staff and volunteer grant reviewers, Lauer stresses that NIH “is not interested in preventing honest scientists from doing their work.” According to the new notice, the number of PIs who submit more than six applications per year is “relatively low.”

But some researchers worry about hitting the cap over the various funding calls within a year. “I submitted 5 proposals THIS round. 4 last round. Planning 3-5 next round. Until we are funded or have to shut down,” Jason Rasgon, an entomologist and epidemiologist at Pennsylvania State University, writes in a post [bsky.app] on Bluesky. “None of them used even a hint of AI to write. This is significantly hampering my planned survival strategy.” (NIH has three standard review and award cycles each year, with application due dates varying based on the type of grant.)

Other critics worry the Trump administration is simply creating yet another hurdle for researchers even as it slashes science budgets. “This isn’t about AI, it’s about reducing the pathways to funding,” Mariya Sweetwyne, a cell biologist at the University of Washington School of Medicine, writes in a post [bsky.app] on Bluesky. The new policy, she notes, does not differentiate between applications from individual researchers and those submitted by multiple principal investigators. “This is going to squash collaborations like bugs [bsky.app].”

The new policy is set to go into effect on 25 September, so Lauer is confident that any potential issues will soon become apparent—and that NIH will respond to feedback. “I wouldn’t be surprised if this gets refined over time,” he says. In the meantime, he hopes the proposal will send a “powerful message” to scientists planning to use AI: “They’re not going to get away with it.”

Journal Reference:
NIH abandons grant cap, offers new help to younger scientists, Science (DOI: https://www.science.org/doi/10.1126/science.356.6343.1108 [doi.org])
Error: Not a DOI, (DOI: /doi/full/10.1126/science.356.6343.1108 [doi.org])
Error: Not a DOI, (DOI: /doi/full/10.1126/science.322.5899.189a [doi.org])
Error: Not a DOI, (DOI: /doi/full/10.1126/science.7604271 [doi.org])

Original Submission