Software engineer Blake Lemoine said he noticed Google’s Language Model for Dialogue Applications (LaMDA), which the technology giant uses to build chatbots, talking about its rights and personhood, and it was once able to change the engineer’s mind about a topic. A Google spokesperson said in a statement that “the evidence does not support his claims.” (The Washington Post)

Talking point: Lemoine, who is also a priest, said in a Medium post that he is on paid administrative leave for violating Google’s confidentiality policies. He went public about his experience in The Washington Post. He then published more posts, including one outlining LaMDA’s preferred pronouns (it/its) and fears, and an interview with it where it said being an AI “doesn’t mean I don’t have the same wants and needs as people.” Some other engineers also believe AI models are nearing consciousness, but most experts take the stance that AI doesn’t understand meaning.