Google AI Defeats World Champion in Complex Board Game

May 4, 2016

Unlocking Word Meanings

Read the following words/expressions found in today’s article. 

1. at the hands of / æt ðə ˈhændz əv / (idiom) – by someone’s actions
Example: She suffered defeat at the hands of her rival.

2. triumph / ˈtraɪ əmf / (v.) – to win
Example: Our national team triumphed in the Olympics.

3. leap / lip / (n.) – figuratively, a great increase or advancement that happened suddenly
Example: Their unexpected victory against the seniors was a great leap for the freshmen.

4. prodigy / ˈprɒd ɪ dʒi / (n.) – someone, especially a young person, who is known for being very talented in something
Example: The chess prodigy won all the matches without difficulty.

5. intuition / ˌɪn tuˈɪʃ ən / (n.) – the ability to feel that something is true or correct even without proof
Example: Her strong intuition allowed her to make good decisions.


Read the text below.
Go world champion Lee Sedol has faced a shocking defeat at the hands of AlphaGo, an Artificial Intelligence (AI) program by Google DeepMind.

Lee engaged the AI in five matches for the Google DeepMind Challenge, held in March in South Korea. Much to the surprise of everyone, AlphaGo won with the score of four wins to one loss.

Prior to beating Lee, AlphaGo also triumphed in October 2015 over Fan Hui, a three-time European Go Champion, with the score of 5-0. However, AlphaGo’s victory against Lee was considered a major leap for AIs because of the Go world champion’s reputation as a prodigy, as well as the difficulty of the game itself.

Go is an ancient Chinese game known for its complexity. The game is played on a 19x19 board, with 361 flat stones colored either white or black. A player wins by capturing opponent’s stones and covering a larger “territory” on the board.

Before AlphaGo, AI developers had difficulty making programs that can play Go skillfully because the game requires intuition. Even experts expected AlphaGo to take 10 more years to be able to beat a professional, but AlphaGo triumphed through its use of deep learning.

Deep learning involves analyzing massive amounts of data to learn a task. The program then trains itself and improves by learning from its mistakes. AlphaGo analyzed 30 million moves from expert Go players to learn the game, and then played Go with itself millions of times. AlphaGo developers say it can correctly predict moves of opponents 57% of the time, and is expected to further improve after the match with Lee.

Viewpoint Discussion

Enjoy a discussion with your tutor.  

Discussion A

·         Do you think that humans should be scared of advanced AI programs like AlphaGo? Why or why not?
·         In what ways can AI programs that can learn on their own be useful? Discuss possible applications.

Discussion B

·         How do you think would a rise in the use of Artificial Intelligence affect society?
·         What kind of jobs can AI programs possibly replace in the future? 

May 4, 2016