DOWNLOAD VIDEO
Could you beat a machine at a video game? With a new system created by artificial intelligence researchers, it just became less likely.
Researchers from the artificial intelligence company DeepMind Technologies have invented a system that can play games from the Atari 2600, a gaming console that was released in 1977. The system is one of the first artificial intelligence technologies to demonstrate the ability to learn,
according
to technology news website Ars Technica.
The system operates on a series of computer algorithms that simulate the learning process, enabling it to learn how to play the games independently. It demonstrates a “general nontrivial intelligence,” a term researchers use to describe a system that can perform some of the same tasks as humans.
It can now play 49 Atari games, and it plays half the games better than humans, according to DeepMind Technologies vice president Demis Hassabis.
“We don’t actually give any clues to the system about what it is supposed to do in the game…It has to learn all those things from first principles,” Hassabis said.
The concept of artificial intelligence has existed for hundreds of years, but research in the field gained traction during the 1950s. The term “artificial intelligence” first emerged in 1956 as part of a conference at Dartmouth College, which helped establish the discipline.
Artificial intelligence may eventually bring risks, but systems are not currently developed enough to be a cause for concern, Hassabis said.
Warm up questions
- What is artificial intelligence?
- What is “learning”? How do you learn new information?
- What makes artificial intelligence different than the computers you have used?
Critical thinking questions
- How have you seen artificial intelligence portrayed in movies or other media? How is this new technology similar or different to how you normally think of artificial intelligence?
- Why would researchers want to develop artificially intelligent systems? What are some potential uses of these systems?
- What could be some of the “risks” that Hassabis mentioned of developing artificial intelligence?
- Whose responsibility is it to guard against the dangers of artificial intelligence?
