Researchers Choose Human Skills Over AI Shortcuts

Marine zoologist Danielle Crowley often feels like the only person at the party who doesn't drink. In her PhD cohort at Bangor University in the UK, she's virtually alone in refusing to use generative AI, even as colleagues and lecturers encourage her to try it for everything from coding to conference posters.

But Crowley isn't anti-progress. She's pro-mastery.

"Coding is a skill I want to learn and develop, because it's not the thing I'm the most confident in," she explains. She'd rather struggle through mistakes and actually build her abilities than let AI do the heavy lifting.

She's part of a surprising trend. While AI use among researchers jumped from 37% to 58% in just one year according to an Elsevier survey, a significant group of scientists are deliberately opting out.

Their reasons go beyond personal learning curves. Hugh Possingham, a conservation scientist at the University of Queensland, has witnessed AI "hallucinations" that would be funny if they weren't so dangerous to scientific accuracy. He once read a master's thesis citing a researcher who had died ten years before the referenced paper was published.

Chemist Audrey Moores at McGill University spotted AI-generated molecular structures that were simply impossible. "It's like you're asking a three-year-old to draw a chemical," she says. The models produce confident-looking diagrams of molecules that couldn't exist in reality.

Social scientist Tanisha Jowsey discovered an ironic twist. As her faculty's designated "AI champion," she's supposed to promote these tools. Instead, she found that verifying AI output takes longer than doing the work herself 95% of the time.

Why This Inspires

These researchers aren't rejecting innovation. They're protecting something more valuable: the deep expertise that comes from doing hard things yourself, the accuracy that matters in science, and the judgment to know when a shortcut actually slows you down.

Their resistance reminds us that efficiency isn't everything. Sometimes the struggle to learn a skill, write our own words, or solve our own problems is exactly what builds the capabilities we need most.

In a world rushing to automate everything, choosing the human path takes courage.