AI enters the grant game, picking winners
-
In March, Joanna Sadler, an organic chemist at the University of Edinburgh, received an unusual email. It promised £35,000 to advance her research—no strings attached. The offer came from a program at Imperial College London that had developed an artificial intelligence (AI) tool to spot promising climate-related research that was close to commercialization.
Sadler’s work on engineering bacteria to convert disposable cutlery into acetone—an industrial solvent currently derived from fossil fuels—had made the cut. At first, Sadler didn’t really understand what the email was. But after meeting with members from Imperial’s Climate Solutions Catalyst (CSC) team, she realized it was legitimate and decided to take them up on the offer.
As AI continues to improve, universities, public funding agencies, and venture capitalists are all considering how machines could be used to identify breakthrough research. In addition to speeding up grant reviews, some researchers say the technology could help level the playing field for researchers who may be hesitant to seek out commercialization opportunities. But others warn that relying on AI for funding decisions could introduce biases and compromise confidentiality.
-
S Science shared this topic on
-
In March, Joanna Sadler, an organic chemist at the University of Edinburgh, received an unusual email. It promised £35,000 to advance her research—no strings attached. The offer came from a program at Imperial College London that had developed an artificial intelligence (AI) tool to spot promising climate-related research that was close to commercialization.
Sadler’s work on engineering bacteria to convert disposable cutlery into acetone—an industrial solvent currently derived from fossil fuels—had made the cut. At first, Sadler didn’t really understand what the email was. But after meeting with members from Imperial’s Climate Solutions Catalyst (CSC) team, she realized it was legitimate and decided to take them up on the offer.
As AI continues to improve, universities, public funding agencies, and venture capitalists are all considering how machines could be used to identify breakthrough research. In addition to speeding up grant reviews, some researchers say the technology could help level the playing field for researchers who may be hesitant to seek out commercialization opportunities. But others warn that relying on AI for funding decisions could introduce biases and compromise confidentiality.
What cruel joke this all is ughhh.
-
In March, Joanna Sadler, an organic chemist at the University of Edinburgh, received an unusual email. It promised £35,000 to advance her research—no strings attached. The offer came from a program at Imperial College London that had developed an artificial intelligence (AI) tool to spot promising climate-related research that was close to commercialization.
Sadler’s work on engineering bacteria to convert disposable cutlery into acetone—an industrial solvent currently derived from fossil fuels—had made the cut. At first, Sadler didn’t really understand what the email was. But after meeting with members from Imperial’s Climate Solutions Catalyst (CSC) team, she realized it was legitimate and decided to take them up on the offer.
As AI continues to improve, universities, public funding agencies, and venture capitalists are all considering how machines could be used to identify breakthrough research. In addition to speeding up grant reviews, some researchers say the technology could help level the playing field for researchers who may be hesitant to seek out commercialization opportunities. But others warn that relying on AI for funding decisions could introduce biases and compromise confidentiality.
Genuinely I think the world would have much better scientific outcomes if a decent portion of grants were just straight-up randomly allocated (from a pool that meets a minimum standard I guess). Everyone wants to fund ‘almost complete’ research but there’s such little money for the early-stage ‘high risk high reward’ basic research, which is what basically all applied research is built upon.
