Google Quietly Announces The First Self-Learning Artificial Intelligence Agent


This could be one of the biggest stories in the history of humanity, but the first self-learning agent received little publicity from the mainstream media.

Scientists with Google have created the program “Agent,” which has already learned how to play dozens of different computer games, a first in the world of artificial intelligence.

The research was conducted by a British company called DeepMind, who was purchased by Google last year for £400m.

“This is the first significant rung of the ladder towards proving a general learning system can work. It can work on a challenging task that even humans find difficult. It’s the very first baby step towards that grander goal … but an important one,” Demis Hassabis, the company’s founder said.

“With Deep Blue, it was team of programmers and grand masters that distilled the knowledge into a program. We’ve built algorithms that learn from the ground up,” he said.

A recent study published in the journal Nature, found that the agent performed at 75% or better than a professional game tester would on half of the games tested.

Below is a video documenting how “Agent” has taught itself to play video games, and even learn winning strategies:

John Vibes writes for True Activist and is an author, researcher and investigative journalist who takes a special interest in the counter culture and the drug war. 

To Top