泥中蟠龍's Game愛歌
[A love song for games of the dragon waiting for an opportunity]
Artificial intelligence, AlphaGo, and Space Invaders
Technology seems less relevant to a video game. A language created and used by artificial intelligence (AI) was reported on the news recently. To sums up, machine intelligence, which was in the process of learning how to negotiate with the other AI, made the new language for itself to bring about better results. The method of AI communication was so quite unlike any other existing human languages that nobody could understand their conversation.
The brand-new language can imply many meanings. In the first place, human languages might not be effective. Negotiators usually exchange a lot of opinions to find a way to cope with a problem. AI may have made the entirely-new one which can replace inefficient human words. Secondly, technology on artificial intelligence has already developed enough to create a new language. Inventing a new language is a high level of mental activity. There have been no species who communicate well using their languages apart from the human. But, what was surprising to me was that this case may mean that AI can out of human control soon.
I was worried about it when Go master Lee Se-dol played against AlphaGo, a famed AI creation of Google. In the match with Lee, Alphago placed its unexpected moves and not only the professional Go player but also countless viewers from all over the world couldn't figure out what they mean. AI's unpredictable behavior would frighten us. AI technology has been growing and expanding in different areas of our lives. Nowadays, the Internet of things (IoT) technology controls home appliances such as television, refrigerator, vacuum cleaner, or washing machines, etc. via devices connected to a network. If machine intelligence continues to grow without control, humans may have to live in an unstable situation where our household appliances supervise the home environment talking to one another with indecipherable words.
Isaac Asimov, who was known for his works of science fiction and pop-science, introduced "Three laws of robotics" in his short story. First, a robot may not injure a human being or, through inaction, allow a human being to come to harm. Second, a robot must obey the orders given it by human beings except where such orders would conflict with the First Law. Third, a robot must protect its existence as long as such protection does not conflict with the First or Second Laws. Some experts say that the Fourth law needs to be added that a robot has to explain to a human why it is doing something. Understanding of the process would probably help human progress because any powerful technology is inherently double-edged.
Artificial intelligence technologies are variously applied to the video game industry and I want as many as game developers to take AI-related problems very seriously. What AlphaGo's development team made the AlphaGo at first was playing a classic arcade game Space Invaders, which may signify that developments in AI are closely related to the video games.
※ This is from Kyunghyang Games column by 泥中蟠龍 since September 2013.
(http://www.khgames.co.kr)
Translation by Kim Ki-hui
No comments:
Post a Comment