Skip to main content

All hail our robot overlords: Google’s new AI will beat you at Atari

google ai atari wargames
Image used with permission by copyright holder
Researchers at Google have created an artificial intelligence that teaches itself how to play video games. Representatives called it “the first significant rung of the ladder” to developing broad, flexible, intelligent systems, Bloomberg reports.

Google acquired the London-based AI startup DeepMind, which helmed the project, early in 2014.  The program was set to play 49 games on the Atari 2600 console with no instructions. Left to its own devices, the AI was able to best expert human players in 29 of the games, and outperformed the best known algorithms for 43.

Recommended Videos

Rather than programming in particular strategies for each game, the researchers paired a general AI with a memory and a reward system. After completing a level or achieving a high score the system is rewarded, encouraging it to replicate what worked. With clear, measurable goals and the ability to refer to its own memory and adjust behavior based on what happens, Google’s system is able to train itself without human supervision. That is a major leap ahead of other machine learning projects that generally require people to provide feedback. You can dig a little deeper into the tech and see the improvement in action for Breakout on the Google Research Blog.

Demis Hassabis, a DeepMind co-founder and now VP of engineering at Google, said that this is the “first time anyone has built a single learning system that can learn directly from experience and manage a wide range of challenging tasks.” Creating intelligent systems that can navigate unexpected circumstances, rather than simply performing prescribed tasks, is a holy grail for AI research, and this is a major step toward that goal.

Games provide a perfect framework for training these types of flexible AIs for the same reasons that they can serve as a useful pedagogical tool for humans. As simulations, games can recreate the noisy and dynamic environments that an intelligence may have to deal with, but structuring that complexity into manageable systems and discrete goals. The researchers described games’ utility to the journal Nature as a way to “derive efficient representations of the environment from high-dimensional sensory inputs, and use these to generalize past experience to new situations.”

Having mastered the 2D worlds of ’80s Atari games, Hassabis’ plan is to move up into the 3D games of the ’90s. He is particularly excited about driving games, because of Google’s vested interest in self-driving cars. “If this can drive the car in a racing game, then potentially, with a few real tweaks, it should be able to drive a real car. That’s the ultimate aim.”

Google is one of the leading hubs of artificial intelligence research in the world today. In addition to acquiring DeepMind, over the last few years the internet search giant has invested millions in other AI startups and a partnership with Oxford University. Hiring inventor, futurist, and chief Singularity proselytizer Ray Kurzweil as director of engineering in 2012 sent a clear signal about the company’s blue sky ambitions for the disruptive future of artificial intelligence.

The concept for DeepMind’s gaming program is remarkably similar to the 1983 film WarGames, in which an AI called WOPR (War Operation Plan Reponse), designed to provide strategic oversight to the American nuclear missile defense, was trained by playing games like chess and tic-tac-toe. WOPR almost turned the Cold War into World War 3 due to a misunderstanding, so let’s hope that Google and the government never decide to weaponize the project, or at least teach it about no-win scenarios first.

In any case, it is only a matter of time now before the King of Kong is dethroned by a machine in a Kasparov/Deep Blue kind of situation. Is nothing sacred?

Will Fulton
Former Digital Trends Contributor
Will Fulton is a New York-based writer and theater-maker. In 2011 he co-founded mythic theater company AntiMatter Collective…
New Atari 50 DLC shows the Intellivision acquisition is already paying off
An Atari 2600+ sits on a table.

Digital Eclipse's Atari 50: The Anniversary Celebration is an excellent and comprehensive look back at the company's now classic video game lineup, with games to play and extra content to interact with. So far, it's gotten one DLC: The Wider World of Atari, that added even more titles. Now, it's about to get its second, thanks to an acquisition it made earlier this year.

Atari announced The First Console War on Friday, and it's about, as you can guess, the company's first console war with the Intellivision, although it'll touch on a specific element of it. In the 1980s, Mattel was publishing games on the Intellivision. At some point, it decided to release versions of these console exclusives for its main competitor, the Atari 2600, under the M Network label. There are 19 of these games coming to Atari 50 with The First Console War, which is set to launch on November 8 for PC, Nintendo Switch, Xbox Series X/S, Xbox One, PlayStation 5, and PlayStation 4 Atari 50 owners.

Read more
ChatGPT Search is here to battle both Google and Perplexity
The ChatGPT Search icon on the prompt window

ChatGPT is receiving its second new search feature of the week, the company announced on Thursday. Dubbed ChatGPT Search, this tool will deliver real-time data from the internet in response to your chat prompts.

ChatGPT Search appears to be both OpenAI's answer to Perplexity and a shot across Google's bow.

Read more
Google expands AI Overviews to over 100 more countries
AI Overviews being shown in Google Search.

Google's AI Overview is coming to a search results page near you, whether you want it to or not. The company announced on Monday that it is expanding the AI feature to more than 100 countries around the world.

Google debuted AI Overview, which uses generative AI to summarize the key points of your search topic and display that information at the top of the results page, to mixed reviews in May before subsequently expanding the program in August. Monday's roll-out sees the feature made available in seven languages — English, Hindi, Indonesian, Japanese, Korean, Portuguese, and Spanish — to users in more than 100 nations (you can find a full list of covered countries here)

Read more