The former Google chief executive says robots will not kill humans in the next 10 or 20 years.
Don't worry, fellow humans, robots are unlikely to enslave or destroy humans for at least a decade or so. Former Google CEO Eric Schmidt (Eric Schmidt) in Munich Security Conference last week (Munich Security Conference) speaking, to the end of the film and television works of popular scene played very lightly, he said:
"Everyone immediately wants to talk about the doomsday scenarios described in all the movies, and I can tell you with confidence that this is at least a decade or two away. We can worry about it, but let's worry about it for a while.
In fact, it is not comforting. In recent years, the rapid development of artificial intelligence and robotics has made such discussions more extensive. The science fiction film about the robot uprising has grown up with a whole generation, so it is almost certain that these scenes are what we imagine when we think about the future of robots.
This is reasonable, of course, but it is unlikely to happen. Schmidt continued:
"The other thing I want to remind you is that there are still serious errors in these technologies, and they should not be used for key decisions in life. So I don't want to sit on a plane driven by a computer. The technology is not reliable enough, and there are plenty of mistakes. Artificial intelligence is only suggestive, it can make people smarter, but I won't let it command and control.
The last sentence is the key. Even though most of us don't think so, the researchers understand that while artificial intelligence can enhance human capacity, it is not yet appropriate to replace humans. The human brain is complex. Although a 40-year-old can't remember wikipedia or beat the best poker players, the typical robot can't handle the simple improvisation that humans excel at.
In fact, most of the work that artificial intelligence and robotics are good at is lower-level tasks, simple and repeatable targets that are easily defined and measured. Robots are not good at improvisation; They require a clear set of rules that will increasingly require security measures that occur when a failure occurs, such as shutting down the machine when a failure occurs.
Although these technologies will continue to improve, the perception of robots will not appear in the short term. For a robot, if it is dangerous, it must be set as dangerous in advance. So, we're not afraid of robots, but humans who write code.