An engineer of Google announced the discovery of "Shocking the World", but was suspended

Author: Geng Zhi

Artificial intelligence is really "alive"?

According to the British "Guardian" and the Washington Post, a software engineer from Google's Google Company announced a amazing discovery of the world's onlookers last week, saying that a Google company's artificial intelligence algorithm was established to build The intelligent chat robot program has "autonomous emotions".

This software engineer, named Blake Lemoine, has worked in Google for seven years and has participated in projects such as legal legal and artificial intelligence. But because of a Christian family in southern United States, he is particularly interested in ethical issues of algorithms and artificial intelligence. So he later worked in Google's artificial intelligence ethics (Responsible AI) team.

Last fall, he began to participate in the research work of Google's Lamda intelligent chat robot program. The main content is to determine whether this smart chat robot program has discriminatory or hatred through interaction with the program, and to do it in time in time Amended.

But when Lemin talked with Lamda about religious issues, he was surprised to find that the artificial intelligence program began to talk about his rights and personality, and even afraid of being closed. 7, 8 -year -old children's "autonomous emotions".

As a result, Lemin quickly reported his discovery to Google's leadership, but the latter believed that Leimin's discovery was not reliable. Other researchers in Google also analyzed the evidence that Leimin came out and believed that there was no evidence to prove that Lamda had autonomous emotions, but there were many evidence to prove that the procedure had no autonomous emotions.

Kelimin did not think that his discovery was wrong, and even sent a period of interior chat with Lamda directly to the Internet, and asked the media to expose the matter.

This chat record quickly attracted researchers all over the world, because from this record, this Google intelligent chat robot program did treat themselves as "people", and it seemed to show "autonomous emotions". With myself with myself "Treasure".

However, Google and other artificial intelligence experts, and even experts in human psychological cognition believe that the reason why LAMDA can conduct such dialogue is because its system is actually the encyclopedia, message boards, and forums on the Internet, etc. A large amount of information that people communicate on the Internet are collected and then imitated through "deep learning". But this does not mean that this program understands the meaning of these contents of people.

In other words, they think that Google's artificial intelligence chat robot only uses the language and information used in imitating or re -reading in people, not the meaning of understanding these languages ​​and information like humans, so it is not possible to say that artificial intelligence has it. "Autonomous emotions".

(The picture shows the cognitive psychology expert of Harvard University Stephen Pin Ke also believes that Mryhenda has made a mistake, and the artificial intelligence is simply imitated by artificial intelligence as an artificial intelligence.)

At present, Mrylande has been suspended by Google because of his series of behaviors. And according to Google, in addition to violating the company's confidentiality agreement to send the internal interactive information with Lamda to the Internet without authorization, he also includes to ask a lawyer to ask this artificial intelligence robot program, and to find the US Congress members of Google " "Inhuman" behavior.

Mleurin said in a group email inside a company that he hoped that other colleagues would take care of him for his "lovely child" for him.

- END -