Drug discovery databases just grew 90% bigger, thanks to AI pulling data from old papers

Researchers developed an AI system that can automatically extract complex drug data from scientific papers. This system expanded existing drug databases for targeted protein degradation by up to 92%, making a lot more information available for drug discovery.

Drug discovery relies on large, structured datasets to train AI models. For years, much of this critical data was buried in scientific publications, making it slow and expensive to access. This new AI workflow makes it much faster and cheaper to build comprehensive datasets, potentially accelerating the discovery of new drugs. It means drug researchers can now train AI models on a significantly larger and richer set of information than was previously feasible.