CRANE, a UK cybersecurity research network, used AI to desk-reject 50% of proposals before human review, including one on the risks of trusting AI models in voice analysis from Lancaster University forensic linguist Claire Hardaker. This seems like AI doing the right thing in protecting itself. 🙃 This is a joke. It’s horrifying. https://www.science.org/content/article/u-k-funder-uses-ai-triage-reject-grant-proposals-leaving-researchers-dismayed?referrer=https%3A%2F%2Fwww.science.org%2Fcontent%2Farticle%2Fu-k-funder-uses-ai-triage-reject-grant-proposals-leaving-researchers-dismayed%3Freferrer%3Dhttps%253A%252F%252Fgo.bsky.app%252F