AI Is Already Replacing Entry-Level Scientists, Researchers Warn

When Hannah Wayment-Steele started her computational biology lab at the University of Wisconsin–Madison, she faced a decision that would have been unthinkable just five years ago. Hiring a research programmer once seemed essential. Today, she isn’t so sure. “I really don’t see a need for that,” she admits, because artificial intelligence can now handle even complex coding tasks.

“Obsolescence of some basic roles in areas such as computer modelling is not even in the future. It’s happening now, because AI is doing this much better than entry-level scientists.” — Xuanhe Zhao, MIT

The Disruption Is Already Here

A comprehensive investigation by Nature has revealed that AI’s impact on scientific employment is no longer theoretical—it’s happening now. The publication spoke with more than four dozen researchers across academia and industry, and the consensus is stark: artificial intelligence is systematically reducing demand for human researchers who specialize in coding and basic data analysis.

The roles most at risk are precisely those that have traditionally served as stepping stones into scientific careers. Graduate students, postdoctoral researchers, and technical staff who once built their expertise through hands-on coding and data work are finding that AI systems can perform these tasks faster and, in many cases, more accurately than humans.

Research programmers—specialists who write code packages that other scientists use—are seeing their positions eliminated. Brian Hie, a computational biologist at Stanford University, puts it bluntly: such jobs “are now obsolete.” The same applies to positions focused on creating simulations and analyzing datasets, according to Xuanhe Zhao, a mechanical engineer at MIT.

The Pipeline Problem

The implications extend beyond immediate job losses. Some scientists warn of a potential collapse in the scientific talent pipeline if entry-level positions continue to disappear. These roles have historically provided crucial training grounds where young researchers develop the skills and intuition necessary for more advanced work.

“You might temporarily get more research per dollar, but the cost would be a collapse of your pipeline and long-term decline.” — Claus Wilke, University of Texas at Austin

Nanshu Lu, a materials engineer at the University of Texas at Austin, confirms that her lab has become “much more conservative in hiring future graduate research assistants and postdoctoral researchers.” While funding uncertainties play a role, she acknowledges that “AI, for sure” is a significant factor in these decisions.

The impact isn’t limited to traditional laboratory roles. Workers in science-adjacent fields are also feeling the pressure. The American Translators Association reports that membership in their Science & Technology Division has declined by 26% in less than two and a half years, as AI-powered translation tools have improved dramatically.

What Remains Human

Despite these disruptions, researchers identify several categories of work that remain resistant—at least for now—to AI replacement. Positions involving hands-on experimentation are considered safer, as are the roles of senior scientists who organize and coordinate complex research projects.

Creative and strategic thinking also appears to be holding the line. Jonathan Oppenheim, a quantum physicist at University College London, uses AI to generate mock peer-review reports of his manuscripts before submission. He finds the critiques helpful, but notes a critical limitation: AI “is not able to really come up with novel ideas.”

Anton Korinek, an economist at the University of Virginia, predicts that jobs involving “purely cognitive tasks will be first” to be automated. “Traditionally, these are the jobs that were most closely associated with scientific research,” he observes. “They will shortly be taken over by AI.”

The Bigger Questions

The transformation raises fundamental questions about how scientific training should evolve. If entry-level coding and data analysis are increasingly handled by AI, what skills should aspiring researchers prioritize? How will the next generation of scientists develop the intuition that comes from wrestling with complex datasets and debugging intricate code?

For now, the scientific community is grappling with these questions in real time. Some researchers are adapting by focusing on skills that complement AI capabilities—interpreting results, designing experiments, and asking the right questions. Others are racing to master the AI tools themselves, hoping to stay ahead of the curve.

What is clear is that the era of AI as a mere laboratory assistant has passed. It is now a competitor for the very positions that have long formed the foundation of scientific careers. The full implications of this shift will unfold over the coming years, but the direction is unmistakable.

This article was reported by the ArtificialDaily editorial team. For more information, visit Nature.