How AI may speed up drug development

Machines may be better than humans at a critical part of the drug-discovery process — and billions of dollars are riding on their abilities.

Driving the news: Google's DeepMind recently "beat seasoned biologists at predicting the shapes of proteins, the basic building blocks of disease," Bloomberg reports. Startups, big biotech companies and venture capital firms are all starting to take notice.

Why it matters: AI gets thrown around a lot in health care, and a lot of it's empty hype, but some kind of machine learning could have real potential to make the drug-development process faster and remove some of the trial and error, experts told Bloomberg.

• A handful of drugs that have relied on early AI tools are already in human trials, and experts said identifying the best candidates for advanced trials may be one of the technology's biggest near-term capabilities.

By the numbers: AI startups with a focus on biotech pulled in more than $1 billion from VCs last year, according to Bloomberg. That's 4 times more than VCs invested just 2 years earlier.

Go deeper: How Google is a step ahead in the race to build health care AI