Google pays more than $500 million for artificial intelligence startup DeepMind

January 27, 2014

DeepMind Technologies logo

Google bought DeepMind Technologies, a London-based startup working in machine learning and systems neuroscience.

Before the deal, DeepMind was, according to sources cited by Re/code, “the last large independent company with a strong focus on artificial intelligence”, competing with Google, Facebook and other Internet giants for AI talent. The three-year old company was founded by Demis Hassabis, a neuroscientist, AI researcher, and former child-prodigy chess player, Shane Legg and Mustafa Suleyman. The startup secured over $50 million in funding from several tech luminaries and had about 50 employees, according to Re/code, who first reported on the deal.

Google confirmed the deal to several publications, but did not comment on the price.

According to The Information, Google paid in excess of $500 for DeepMind, while Re/code puts the figure at $400 million. Regardless of the cost, Google is said to have competed with Facebook for the startup – DeepMind and Facebook were in advanced talks, but the deal fell through late last year.

DeepMind asked Google to establish an AI ethics board

One of the conditions that DeepMind raised was for Google to establish an “ethics board” that would ensure that artificial intelligence technology is not abused. It’s not clear whether this board would supervise the use of DeepMind technology or of Google’s AI efforts in general.

The deal was reportedly spearheaded by Google CEO Larry Page himself. If Hassabis and DeepMind’s other talent will work for Google, they will join other AI experts including Anna Patterson and Jeffrey Dean. Futurologist Ray Kurzweil is also working at Google with the stated goal to create a search engine that acts much like a “cybernetic friend”.

This is how DeepMind describes itself on its one-page website:

We combine the best techniques from machine learning and systems neuroscience to build powerful general-purpose learning algorithms. Our first commercial applications are in simulations, e-commerce and games.

Google is investing heavily in machine learning, artificial intelligence, and robotics. Under the leadership of former Android boss Andy Rubin, the company recently acquired several robotics company, including the high-profile Boston Dynamics.

Machine learning already has promising applications in Google’s main business, search. Google is training its powerful algorithms in fields like image recognition and voice recognition, and its long-term goal is incredibly ambitious – to build a Star Trek-like computer. Any intelligence artificial breakthrough will eventually impact Android as well. Google Now is already surprising users with its predictive capabilities, and that’s likely just a taste of what’s to come.