AI churns out novel drug candidates in 46 days

It wasn’t that long ago that in silico drug development had its fair share of skeptics. But the artificial intelligence approach to finding new drug candidates is gaining both steam and esteem: Take this new paper in Nature Biotechnology, which shows how an AI tool can generate compounds able to hit a specific disease target — in only 46 days.

The tool, called a generative adversarial network, has most famously been used to create eerily realistic fake photos of celebrities. Many drug companies are using these GANs to try to speed up their R&D processes — though, of course, none of these efforts has yet resulted in an FDA approval.