| AI vs. Human |
August 29, 2019
|
| He browsed in a bookstore and happened to see this book. Recently, he didn't buy new books often because they were expensive. This bookstore had seats for customers, so he usually read parts of some books, but didn't buy them. He noticed that recent books had larger letters and broader space between the lines. Maybe it was for readers who were busy or just lazy to read the conclusion easily. But this time, after reading this particular book for 20 minutes or so, he decided to buy it. The title of the book was, "AI vs. Children Who Can't Read Textbooks". He had read somewhere that the author and her group had researched about AI and came to the conclusion that AI wouldn't replace human beings in the future, so there wouldn't be a 'singularity'. He was very curious about AI and 'singularity'. Her writing was easy to understand, organized and logical. Maybe it was because she was a mathematician and a professor of mathematics. From several years ago, 'singularity' had been feared by many people. But she believed that a 'singularity' world wouldn't come from her specialty. But giants of the industry, like Stephan Hawkins and Bill Gates, warned about 'singularity'. So she started her project with some others including AI engineers and researchers in order to prove that "an AI robot could NOT enter Tokyo University".  When she asked for professional insights from engineers in big computer companies in the US, they laughed, saying it was impossible. It seemed like professionals knew that AI wasn't as smart as a human even after an AI robot defeated a GO master.AI computers' method of 'deep learning' was just using a collection of bid data. From this method, they could get better ways, but it was totally different from "they are thinking". For example, in translation by AI, AI doesn't translate by 'understanding' the sentences or words, they just calculate the combination of words and make the sentence that most people would believe is normal or suitable. AI translation via provability rather than translation via understanding. As they don't understand the sentences' or words' meanings, sometimes mistakes happen that people normally don't make, like, "A mouse is chasing a cat." or "A stone loved the sky." . Because of these mistakes, engineers will have to give more instructions and/or gather more big data to correct these mistakes. He once had wondered why computer companies had so many workers. If AI computers thought for themselves, they could find out their mistakes, and correct their mistakes themselves, like what workers do. But, in reality, as AI computers don't really understand their jobs' meanings, so it is impossible for them to correct themselves. They are good at finding the fastest route in a traffic jam. But they don't understand what human beings are, so they can't help human beings to be themselves. The author pointed out problems about AI computers. She agreed that AI computers would take some human jobs. Some people believed that if some jobs were replaced by AI computers, there would be new jobs created. But she said that it wouldn't happen. In past times, cars took away horse and cart drivers' jobs but those drivers could become taxi drivers after training. But what of jobs that are taken away by AI computers? What can't AI computers do? Companies are supposed to pursue ways to increase benefit with as little cost as possible, therefore many workers should lose their jobs. Her team gave up in trying to enter their AI robot into Tokyo University, but they found that their AI computers are smart enough to be able to enter into some good universities. For the exams, the computers couldn't think, but they could calculate the most possible correct answers. AI computers might make mistakes, but human workers have to correct them. No problem. So, many jobs would be replaced by AI computers, even engineers themselves might not be needed as much. The author suggests that we should shift our jobs to things that AI computers were not good at. Only human beings can think, feel, and create things. And when she thought about this, she was concerned about present students' who will be future workers reading comprehension ability. Reading is very important because when you study, you need to read; and, reading makes you think deeply. However, her team found that students' reading comprehension ability might be getting worse not only in this country, but also in other countries. At the last part of the book, she revealed that this book's royalties will be donated to the organization to promote reading comprehension tests for students and analyzing those results. She closed this book's ending saying, "Let's have a happy 2030 together." He was very pleased to buy this book. |
| |
*conclusion :結論
*specialty :専門
*via :による
*be supposed to :することになっている
*comprehension :理解
*reveal :明かす
*royalty :印税