1/ Anthropic announced that Claude discovered a new enzyme system with CRISPR-like properties.
2/ The announcement was made public on 2026년 9월 23일, and it shows that AI can be used in life-science research to search for candidates and propose new hypotheses.
3/ This case is meaningful in that large language models can be used not only for literature organization or summarization, but also for discovering research ideas and exploratory analysis.
4/ In particular, it suggests that AI can speed up a researcher’s initial exploration in complex search problems such as discovering biological functional units. [1]
Implications
5/ For materials and science researchers, this shows that AI is moving deeper into exploratory research such as literature mining, reaction network exploration, and discovery of candidate materials, proteins, and enzymes.
6/ However, because verification and reproducibility are essential for such results, it is important to have a system that connects generated hypotheses with experimental design and independent confirmation procedures.
7/ Rather than accepting AI-proposed candidates as they are, a promising approach is to combine them with surrogate models or active learning to improve validation efficiency. [1]