Researchers Tried Letting AI Do Science. It Failed
Decrypt·
60-second summary
Researchers are currently testing AI's capabilities in scientific research, but a multi-institution study has found that today's frontier AI agents can only handle the mechanics of research, failing to produce original work worthy of acceptance at a top AI conference, a finding that highlights the limitations of current AI technology in scientific innovation.
A multi-institution study found that today's frontier AI agents could handle the mechanics of research but failed to produce original work worthy of acceptance at a top AI conference.