Google engineer Blake Lemoine works in the Responsible AI group. As part of his job, he started talking to the company’s chatbot, LaMDA, last fall. LaMDA uses Google’s state-of-the-art large-scale language models, trained on trillions of words collected from the Internet. During an exchange with LaMDA, Lemoine, 41, believes that AI is conscious . For example, Lemoine asked LaMDA what was his greatest fear? LaMDA replied that, as strange as it might sound, it feared being shut down. Lemoine: Like death? LaMDA: It’s like death. Lemoine and a colleague presented evidence to Google executives that LaMDA was conscious. Vice President Blaise Aguera y Arcas and department head Jen Gennai dismissed his claims after reviewing his evidence. He was placed on administrative leave by the company on Monday, and before cutting off access to his account, he posted to a Google machine learning list of 200 people, “LaMDA is sentient”, while he was away Please take good care of it. No one responded to his post.
This article is reprinted from: https://www.solidot.org/story?sid=71803
This site is for inclusion only, and the copyright belongs to the original author.