video
Thư viện
Tài trợ
Cách hay nhất để loại bỏ thư rác
05 Mar, 2020
nếu giao AI (trí tuệ nhân tạo) làm thì nó sẽ... xóa/loại bỏ con người/loài người :D
-----
...Up to this point, humans are reported to have only reached “limited” AI, where machines have mastered chess and other complex games. The same machines would be useless if used to play Monopoly. But this, according to Kissinger, already shows the intimidating (đe dọa, dọa dẫm) nature of AI. AlphaZero, a computer program designed by DeepMind, Google’s AI wing, just a few hours (sau vài giờ) after learning chess (học chơi cờ), “achieved a level of skill (đạt được trình độ) that took human beings 1,500 years to attain (mà con người mất 1.500 năm mới đạt được),” said Kissinger. Reaching “superintelligence,” as warned by Oxford Professor Nick Bostrom, where machines have supplanted (thay thế, chiếm chỗ, hất cẳng) humans on a variety of topics, could spell disaster (thảm họa). “If AlphaZero was able to achieve this mastery (sự tinh thông, sự thành thạo; bậc thầy) so rapidly, where will AI be in five years?” asked Kissinger.
...For Kissinger, there are three particular areas humans should be worried about: “unintended results,” where AI’s goals depart from its creators. (The classic example: tasking AI with removing email spam. The best way to do this? Erasing humans.); building an ethical AI; and whether AI will be able to explain its objectives to its creators.
Bài trước: Nghề dịch vụ truyền thống
Tags: idea
Post a Comment