X

Google's New AI Predicts Molecular Behavior, Which Could Supercharge Drug Discovery

AlphaFold 3 can help scientists design more effective drugs, Google says.

Imad Khan Senior Reporter
Imad is a senior reporter covering Google and internet culture. Hailing from Texas, Imad started his journalism career in 2013 and has amassed bylines with The New York Times, The Washington Post, ESPN, Tom's Guide and Wired, among others.
Expertise Google | Internet Culture
Imad Khan
2 min read
Google DeepMind

Google DeepMind at Google I/O 2023. 

James Martin/CNET

Google DeepMind, the company's artificial intelligence research lab, has a new molecular prediction model called AlphaFold 3 that it says can predict the interactions of "all life's molecules" with at least 50% improved accuracy over past models. This means scientists could develop drugs that target diseases more effectively, Google said in a blog post and in a paper published in the journal Nature on Wednesday. 

AlphaFold 3, available to researchers for free, will allow scientists to test different structures composed of proteins, DNA and RNA to generate predictions of how different substances might interact. 

AI Atlas art badge tag

Doing this type of predictive modeling with AlphaFold 3 could greatly speed up research while also reducing costs -- Google says that in the past, creating an experimental protein structure prediction would take years and cost hundreds of thousands of dollars. 

Google purchased DeepMind in 2014 for an estimated $400 million to $650 million to expand its AI capabilities. DeepMind made headlines for creating systems that could beat the world's best Gochess and shogi players. It soon started making advancements in protein folding and crystalline structure discovery.

AlphaFold 2 was used to help in malaria and cancer treatments as well as enzyme design, Google says. AlphaFold 3 takes things a step further by also looking at the "broad spectrum of biomolecules," which could aid in the discovery of biorenewable materials to drug design and genomics research.

Google said it's already been used to predict hundreds of millions of structures that would have taken "hundreds of millions of researcher-years at the current rate of experimental structural biology."

Isomorphic Labs, the sister company to DeepMind, is already working with pharmaceutical companies to develop new treatments.

DeepMind's advancements in helping scientific research isn't only good for humanity but also helps give Google a good brand image as a tech company hoping to use its power and expertise to help solve monumental challenges. Advancements at DeepMind can also help Google enter new markets. 

For more on Google's everyday AI tools, be sure to read CNET's hands-on reviews of Gemini and ImageFX, along with reviews of chatbots including ClaudeAdobe Firefly and OpenAI's ChatGPT.

Editors' note: CNET used an AI engine to help create several dozen stories, which are labeled accordingly. The note you're reading is attached to articles that deal substantively with the topic of AI but are created entirely by our expert editors and writers. For more, see our AI policy.