Not like us: Artificial minds we can't understand

 作者:眭泠失     |      日期:2019-03-01 05:20:01
(Image: Robert Hodgin) By Douglas Heaven It can see things we miss and knows us better than we know ourselves Rick Rashid was understandably nervous. As he stepped onto the stage to address 2000 researchers and students in Tianjin, China, he was risking ridicule. He didn’t speak Chinese, and his translator’s poor skills in the past promised embarrassment. “We hope that in a few years we’ll be able to break down the language barriers between people,” the senior vice-president of Microsoft Research told the audience. There was a tense 2-second pause before the translator’s voice came through the speakers. Rashid continued: “Personally, I believe this is going to lead to a better world.” Pause, repeat in Chinese. He smiled. The crowd were applauding every line. Some people even cried. This seemingly overenthusiastic reaction was understandable: Rashid’s translator had come far. Every sentence was understood and delivered flawlessly. And the most impressive part? The translator was not human. Performing such a task was once far beyond the abilities of the most sophisticated artificial intelligence, and not for want of effort. For years, AI was dominated by grand plans to replicate the performance of the human mind. We dreamed of machines that could understand us, recognise us and help us make decisions. In the last few years we have achieved those goals. But not in the way the pioneers imagined. So have we worked out how to replicate human thinking? Far from it. Instead, the founding vision has taken a radically different form. AI is all around you,