Artificial Intelligence, Scientific Discovery, and Product Innovation
2024/12/21 by Aidan Toner-Rodgers, Toner-Rodgers, Aidan · 8 voices · 1 citation
Business, Management and Accounting · #Big Data and Business Intelligence #econ.GN
paper · pdf · doi:10.48550/arxiv.2412.17866
openalex publication_date 2024/12/21 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Abstract
This paper studies the impact of artificial intelligence on innovation, exploiting the randomized introduction of a new materials discovery technology to 1,018 scientists in the R&D lab of a large U.S. firm. AI-assisted researchers discover 44% more materials, resulting in a 39% increase in patent filings and a 17% rise in downstream product innovation. These compounds possess more novel chemical structures and lead to more radical inventions. However, the technology has strikingly disparate effects across the productivity distribution: while the bottom third of scientists see little benefit, the output of top researchers nearly doubles. Investigating the mechanisms behind these results, I show that AI automates 57% of "idea-generation" tasks, reallocating researchers to the new task of evaluating model-produced candidate materials. Top scientists leverage their domain knowledge to prioritize promising AI suggestions, while others waste significant resources testing false positives. Together, these findings demonstrate the potential of AI-augmented research and highlight the complementarity between algorithms and expertise in the innovative process. Survey evidence reveals that these gains come at a cost, however, as 82% of scientists report reduced satisfaction with their work due to decreased creativity and skill underutilization.
Cited by
Discussions
- It blew up when a third party unconnected with MIT asked the very simple question of where the data had come from? There are very few materials companies of a size to fit the description and none had [bsky, 17 points, 1 comments]
- It was cosine similarity of TF-IDF vectors: arxiv.org/abs/2412.17866 Reading the Appendix, there are things I don’t like (lack of details like inter-annotator agreement, too much math-y notation obscu [bsky, 9 points, 3 comments]
- Came across the statement by MIT yesterday that this preprint should be retracted: arxiv.org/abs/2412.17866 [bsky, 4 points, 1 comments]
- Searching on "toner-rodgers arxiv" gives one hit matching the description, link here: arxiv.org/abs/2412.17866 And yes, there should immediately have been questions like "are the 40% more patents in t [bsky, 1 points, 1 comments]
- arxiv.org/abs/2412.17866 Feels like another canard. Sorry to say it. But who the fuck wrote the paper? [bsky, 0 points, 1 comments]
- Very interesting study on the impact of AI as a tool in research and development. Study in materials lab showed that using AI significantly improved research productivity, but that this depended on ho [bsky, 0 points, 1 comments]
- 2) In the domain of materials discovery. LLMs led to more materials, patent filings, and downstream innovation. Most interestingly, diversity between scientists on how much they improved - scientific [bsky, 0 points, 1 comments]
- arxiv.org/abs/2412.17866 Survey evidence reveals that these gains come at a cost, however, as 82% of scientists report reduced satisfaction with their work due to decreased creativity and skill underu [bsky, 0 points, 1 comments]
Related