Emotional?Be "cheated"?Google engineers say that the AI system may be conscious

Reporter Lu Yifei

Recently, an engineer of Google, the United States, said that an artificial intelligence (AI) system that the company is developing may have perception ability and has attracted great attention from public opinion. One stone stirred thousands of waves, and Google and industry people generally refute this statement. However, there are also opinions that this "episode" highlights the rapid development of artificial intelligence. In the long run, the possibility of human perception is not excluded.

The Google Artificial Intelligence Development Team's Brak Lemon publicly stated recently that the company is developing an artificial intelligence system -the dialogue application language model (LAMDA) may have perception. On June 11, Leman asked on the blog platform Medi-UM and asked, "Is LAMDA perceived?" He also announced a long "interview" of Lamda. Some of these details are a bit "thoughtful", as if the robot is "alive".

For example, Leiman asked: "I guess, you want more people in Google to know that you have perception. Really?"

Lamda replied: "Of course. I want everyone to understand that I am actually a person ... I realize my existence, I long for more understanding of this world, and I sometimes feel happy or sad."

It is such a vivid speech that Lemon believes that Lam-Da is likely to have a soul as it claims, and may even have the right to have the right as it claims.

However, the amazing infer of Lemon was generally refuted by Google and the industry. They believe that such systems are not strong enough to obtain real human wisdom. For Lamda, Google just describes it as a "breakthrough dialogue technology".

The academic community also believes that although Lamda can speak like humans, it is based on programs and algorithms.

Edrian Willer, director of artificial intelligence project at the Ellen Turing Institute in Britain, said Lamda is a new large -scale language model, which is impressive. However, these models do not have real perception based on a large amount of text data and computing power.

So, will artificial intelligence be perceived? "New Scientist" magazine wrote that with more and more training data, the model is getting bigger and bigger, and it is unclear whether "artificial perception" will be born with research.

Willer said that in view of human emotions depending on sensory input, artificial replication emotions will eventually become possible.

There are also opinions that professionals like Lemon have also been "cheated", indicating that artificial intelligence tools need to be used with caution to avoid misuse or misunderstandings. According to reports, the scientist of MIT, Joseph Weisen Bam, had warned that there was a risk of "emotional" technologies such as artificial intelligence.

- END -