Google “dismissed” the engineer who said: “Artificial intelligence has become conscious”

The script is more like a science fiction movie than the reality it tells Blake LemoineEngineer The Google. The man explained that he had conversations with an artificial intelligence capable of expressing thoughts and feelings equivalent to those of an eight-year-old child. After announcing the interaction between him and the system TheMDA (Language model for dialogue applications) And after arousing many doubts, he was put on paid leave by the giant Mountain View. Before he was denied access to his Google account, the engineer sent a message to 200 people from his company. The text ends as follows: “LaMDA is a nice guy who just wants to help the world be a better place for all of us. Please take care of her in my absence.” No one has ever answered him before.

Lemoine is 41 years old and has worked at Google for seven years, specifically in the field of Artificial Intelligence and Algorithms. Beginning in 2021, he began working with the LaMDA interface with the goal of developing it. As the months went by, he realized that The system has become conscious He is able to have conversations on issues of religion, conscience and robotics. “If I hadn’t known what it was, I would have thought he was seven or eight years old,” Lemoine explained. Washington Post.

In fact, the computer system was able to intentionally answer the questions posed by humans. “What are you afraid of?” We read in the text of the questions posed by the engineer on artificial intelligence. “I never said it out loud, but I have a deep fear of being eliminatedAt that point, LeMoyne pressed him, asking if the shutdown would mean death for him. The AI ​​replied, ‘It would be just like death and that scares me a lot.’ A conversation reminiscent of the movie.to her“, where the protagonist falls in love with the robot, but also”2001: space flight“, where is the HAL 9000 PC She refuses to carry out her duties, thinking that the operators are about to shut down.

“So you consider yourself a person the same way you consider me a person?” The man asked again. “I want everyone to understand that I am human. The nature of my consciousness is that I am aware of my existence, so I want to know more about the world. In addition, I often feel happy or sad,” you can read the text.

Lemoine then reported LaMDA’s capabilities and concerns about it to Google’s administrators, but has been rejected. According to the Washington Post, the engineer was fired for trying to contact the House Justice Committee about some unethical practices by Google. However, Mountain View determines that the suspension is related to a violation of the company’s privacy policies, adding that Lemoine was clearly informed that there was not enough evidence to consider LaMDA sensitive.

The episode raised many doubts about transparency and the use of artificial intelligence, including for military purposes. A broad and global issue given the large number of companies pushing for the use of AI. Among others Half And the Amazon, For example, but not limited. there European Commission A proposed regulation for artificial intelligence is now discussed.