AI use in science linked to more novelty and impact, study finds
Research using artificial intelligence produces more novel work with greater scientific impact than studies that don't use AI, according to a study of over 80 million scientific papers.
Researchers from the University of Strasbourg and the European Commission examined papers published between 2005 and 2023 across more than 170 fields. They identified which studies benefited from AI and why those benefits occurred.
The findings suggest reason for optimism about AI's role in accelerating scientific progress. The analysis shows AI doesn't simply speed up existing research methods-it shifts the kinds of questions researchers pursue and how they approach them.
What the data shows
Papers that incorporated AI tools showed measurable differences in two areas: the novelty of their findings and their subsequent impact in the field.
Novelty matters because it indicates researchers are exploring new territory rather than extending established work. Impact matters because it shows other scientists are building on these results.
The breadth of the dataset-spanning nearly two decades and multiple scientific disciplines-suggests the pattern holds across different research contexts, not just in fields where AI naturally fits.
Why this matters for researchers
If you work in scientific research, the findings suggest AI tools can help you ask better questions, not just answer existing ones faster. The data indicates AI amplifies research quality when integrated thoughtfully into your work.
The study also implies that labs and institutions investing in AI capabilities may see returns beyond efficiency gains. Researchers using these tools appear to be producing work that influences their fields more substantially.
Learn more about how AI is being applied across scientific disciplines through AI for Science & Research resources.
Your membership also unlocks:
