Researchers who use AI tools in their science publish more papers and advance more quickly in their careers, but AI-heavy research clusters around the same data-rich problems, narrowing the scope of science as a whole
https://spectrum.ieee.org/ai-science-research-flattens-discovery
5 Comments
Peer reviewed research article: [nature.com/articles/s41586-025-09922-y](http://nature.com/articles/s41586-025-09922-y)
Im not sure that this title really makes sense, a higher focus in one area does not necessarily narrow the “scope” of anything. Scope refers to the total breadth of all topics and subjects.
I understand how they could be related but i feel like it could have been worded better.
It makes sense, when doing scientific search you have to cite sources and when those sources are being funneled by the same algorithm that scraped the info from itself after it was sanitized, the bucket is gonna narrow artificially.
When the only tool you have is a hammer, every problem looks like a nail.
Sure. Researchers who can write twice as fast, or can process data twice as quickly, are going to do things faster. Are they doing it better? That’s a lot harder to tell. But in a “publish or perish” environment, faster *is* better.