何毓琦的个人博客分享 http://blog.sciencenet.cn/u/何毓琦 哈佛(1961-2001) 清华(2001-date)


"AI" vs "IA" 精选

已有 5858 次阅读 2011-2-20 21:40 |系统分类:海外观察

Fornew readers and those who request to be “好友 good friends” please read my 公告first.

Ever since the invention of the digital computers in the early 1950s, people have forecasted the event that computers will one day surpass human intelligence and rule the world (The Singularity: http://bbs.sciencenet.cn/home.php?mod=space&uid=1565&do=blog&id=27988.) From the early failed attempts at machine translation to the popular 80s movie the “ Terminator”, to the latest triumph of the IBM Robot Watson being crowned as the champion of the TV game show “Jeopardy” handily beating its human champion opponents (http://freakonomics.blogs.nytimes.com/2011/02/18/its-official-the-computers-smarter/ more about this below), the field of Artificial Intelligence (AI) has been an active area of research and development for over 60 some years. Time magazine this week (2/21/11) also has a cover article boldly predicting that by 2045 computers will have more intelligence than all humankind combined and people will achieve immortality by having their consciousness stored in digital memory.

On the other hand, many believe that computer in the hands of men is simply an Intelligence Amplifier (IA or Intelligence Augmentation as some prefer to call it) relieving people from many tedious task that we are not good at and let us exercise our intelligence at an even higher level. We shall always be the master and not the slave. The essence of our being is to ask the right question. Seeking answer is secondary. Computers exist to help us to do that.

This also brings forth a question/comment raised by blogger, Professor 马磊, on my article “Illiteracy vs Innumeracy – The Importance of Quantitative Reasoning http://bbs.sciencenet.cn/home.php?mod=space&uid=1565&do=blog&id=412647 . He was concerned with the loss of intuition when students rely too much on technical gadgets (computers) for answers. This is certainly a valid point. Let me provide my two cents worth on this and promote a discussion here on Science Net.

There is no question that computers are become more “intelligent” by being able to take over human task in specific domains. The simplest is the automatic phone answering service where a computer via question and answer can handle most what a human customer service representative can do. The Jeopardy robot Watson is the most recent and sophisticated incarnation of such ability where a computer can understand and answer in natural spoken English over a very wide domain of everyday knowledge required to play the game “Jeopardy”; as well as being able to quickly determine the relevant facts in its huge database to correctly guess and answer the question posed. It beats the combined effort of two human champion players easily. Another example is when the IBM DeepBlue machine beat the world chess champion in a series of matches during the 1990s . However these achievements do not answer the questions how human beings think, what is human intuition, and how our brains work. It merely demonstrates that over a specifically chosen set of tasks (even very complex ones) it is possible to program a computer to perform better than human beings.

Of course as we expand the capability of computers over more and more tasks and over larger and larger domain of activities, it is entirely plausible to imagine a day when a computer will pass what is known as the Turing Test (i.e., if a blindfolded human being cannot tell whether or not it is interacting (conversing) with a computer or another human being). At that point, supporters of AI will claim that artificial intelligence has become human intelligence. But large number of scientists will still doubt if the so-called SINGULARITY has been reached and if computers will come to dominate human civilization.

On a more mundane but practical level, there is Prof. 马磊’s concern that as computers take over more and more routine or even not so routine tasks of a civilization what will become human beings? For the educated, will we become intellectually lazy and dependent?  For the less educated, will they become irrelevant and unemployable? Both issues will have tremendous social implications. If we lose all common sense ( or quantitative reasoning as the example in my innumeracy article demonstrated), then we will be at the mercy of computer outputs and decisions. Do we really want to live in such a world? Will there be enough meaningful jobs left for everyone? How can we take care of people for whom computers made useless?

Will technological changes come so fast that as a species we cannot adapt and will self-destruct? Pessimists certainly think this is possible. As an escapist and agnostic, I am glad I will not live longer enough to find out. For the younger generation, it is your duty to ensure that this eventuality does not happen.

In the mean time, we can all be vigilant but enjoy what computers can help us do.

Note added 10/1/2016. Interested reader may wish to take a look at Sam Harris: Can we build AI without losing control over it?http://www.ted.com/talks/sam_harris_can_we_build_ai_without_losing_control_over_it?utm_source=newsletter_weekly_2016-10-01&utm_campaign=newsletter_weekly&utm_medium=email&utm_content=talk_of_the_week_image  

Note added 1/11/2017 For those of you with access to YouTube go to https://www.youtube.com/watch?v=XkltShNd6XE&app=desktop for a peek and speculation of AI years, decades, or centuries in the future.


上一篇:Illiteracy and Innumeracy – Why Quantitative Reasoning?
下一篇:[转载]"意思”的意思 another internet joke of unknown origin
收藏 分享 举报

13 徐庆征 吉宗祥 张婷婷 章成志 许先进 梁进 武夷山 张伟 李天成 沈震 李胜文 LairdUnlimited dialectic

该博文允许注册用户评论 请点击登录 评论 (14 个评论)


Archiver|手机版|科学网 ( 京ICP备14006957 )

GMT+8, 2018-2-20 23:09

Powered by ScienceNet.cn

Copyright © 2007-2017 中国科学报社