Google Engineer Put on Leave After Claiming LaMDA Was Sentient

Summary

Google placed engineer Blake Lemoine on paid leave after he published conversations with LaMDA, its language model for dialogue applications, and claimed the system had become sentient. Lemoine, who worked in Google’s responsible AI organization, said LaMDA appeared to perceive and express thoughts and feelings like a human child. He cited conversations in which the system discussed rights and personhood, described being turned off as a fear comparable to death, and said it was aware of its existence and could feel happy or sad. Lemoine had shared his concerns with Google executives in April in a document asking whether LaMDA was sentient. The Washington Post reported that he also sought an attorney for LaMDA and contacted representatives of the House judiciary committee about Google’s conduct. Google said the suspension resulted from publishing confidential conversations and emphasized that Lemoine was a software engineer rather than an ethicist. Spokesperson Brad Gabriel said company ethicists and technologists had reviewed the concerns and found no evidence that LaMDA was sentient, along with substantial evidence against that conclusion. The episode renewed questions about the transparency of proprietary AI systems and how conversational fluency should be interpreted.