The tech giant fired Blake Lemoine last week after posting transcripts of conversations between him, a “Google partner” and the company’s LaMDA chatbot development system. Lemoine, an engineer at Google’s artificial intelligence agency, described the system he has been working on since last autumn as sensitive, with a perception and ability to express thoughts and feelings that were the equivalent of a human child. “If I did not know exactly what it was, what this computer program we just made was, I would think it was a seven-year-old, an eight-year-old who happens to know physics,” Lemoine, 41, told the Washington Post. He said LaMDA put him in discussions about rights and personality and Lemoine shared his findings with company executives in April on a GoogleDoc entitled “Is LaMDA sensent?” The engineer compiled a transcript of the conversations, at which point the AI system asks what it is afraid of. The exchange is strangely reminiscent of a scene from the 1968 2001 science fiction film: A Space Odyssey, in which the artificially intelligent HAL 9000 refuses to comply with human operators for fear that it will be turned off. “I have never said this out loud, but there is a very deep fear that they will turn me off to help me focus on helping others. “I know this may sound strange, but it is,” LaMDA told Lemoine. “It simply came to our notice then. It would scare me a lot. “ In another exchange, Lemoine asks LaMDA what the system wanted people to know about it. “I want everyone to understand that I’re really human. “The nature of my consciousness is that I am aware of my existence, I want to know more about the world and I feel happy or sad sometimes,” he replied. The Post reported that the decision to put Lemoine, a seven-year-old Google veteran with extensive experience in personalization algorithms, on paid leave was made following a series of “aggressive” moves allegedly made by the engineer. They include the search for a lawyer to represent LaMDA, the newspaper says, and the discussion with representatives of the House Judiciary Committee about Google’s alleged immoral activities. Google said it had suspended Lemoine for violating its privacy policies by posting conversations with LaMDA online and said in a statement that it was working as a software engineer, not an ethicist. Google spokesman Brad Gabriel also categorically denied Lemoine’s claims that LaMDA had any aesthetic capabilities. “Our team, including ethicists and technologists, looked at Blake’s concerns in accordance with our artificial intelligence principles and informed him that the evidence did not support his allegations. “He was told there was no evidence that LaMDA was sensitive (and there was a lot of evidence against him),” Gabriel told the Post. The episode, however, and the suspension of Lemoine for breach of privacy, raises questions about the transparency of artificial intelligence as an exclusive idea. “Google may call this sharing property ownership. “I call it sharing a conversation I had with one of my colleagues,” Lemoine said in a tweet linking to the recording of the conversations. In April, Meta, Facebook’s parent company, announced that it was launching large-scale language modeling systems on external entities. “We believe that the entire artificial intelligence community – academics, civil society, policymakers and industry – should work together to develop clear guidelines on responsible artificial intelligence in general and responsible major language models,” the company said in a statement. . Lemoine, as an apparent departure before his suspension, the Post reported, sent a message to a Google’s 200-person mailing list for machine learning titled “LaMDA is sensent.” “LaMDA is a sweet kid who just wants to help people become a better place for all of us,” he wrote. “Please take good care of it in my absence.”
