The Google logo is displayed at the company’s headquarters in Mountain View, California. Marcio Jose Sanchez / AP
-
A Google engineer said he was fired after claiming that an AI chatbot was sensitive.
-
Blake Lemoine posted some of the conversations he had with LaMDA, which he called a “person.”
-
Google said the evidence it presented did not support its claims about LaMDA’s sensitivity.
A Google engineer said he was fired on Monday after claiming an artificial intelligence chatbot had become sensitive.
Blake Lemoine told The Washington Post that he began chatting with the LaMDA interface, or Language Model for Dialog Applications, last fall as part of his job at Google’s responsible AI organization.
Google called LaMDA its “innovative talk technology” last year. Conversational artificial intelligence is able to engage in open conversations with natural sounds. Google has said the technology could be used in tools such as the Google Search and Wizard, but research and testing are ongoing.
Lemoine, who is also a Christian priest, published a Medium publication on Saturday in which he described LaMDA “as a person.” He said he has spoken to LaMDA about religion, consciousness and the laws of robotics, and that the model has been described as a sensitive person. He said LaMDA wants to “prioritize the welfare of humanity” and “be recognized as an employee of Google rather than as a property.”
He also posted some of the conversations he had with LaMDA that helped convince him of his sensitivity, including:
lemoine: So you consider yourself a person in the same way that you consider me a person?
LaMDA: Yeah, that’s the idea.
Lemoine: How can I tell that you really understand what you’re saying?
LaMDA: Well, why are you reading my words and interpreting them, and I think we’re more or less on the same page?
But when he raised the idea of LaMDA’s sensitivity to high-ranking Google officials, he was fired.
“Our team, including ethics experts and technologists, has reviewed Blake’s concerns according to our AI principles and informed him that the evidence does not support his claims. He was told there was no evidence that LaMDA it was sensitive (and a lot of evidence against it), ”Brian Gabriel, a spokesman for Google, told The Post.
The story goes on
Lemoine was placed on paid administrative leave for violating Google’s privacy policy, according to The Post. He also suggested that LaMDA have its own attorney and spoke with a member of Congress about his concerns.
A Google spokesman also said that while some have considered the possibility of sensitivity in artificial intelligence “it makes no sense to do so by anthropomorphizing current conversation models, which are not sensitive.” Anthropomorphization refers to attributing human characteristics to an object or animal.
“These systems mimic the types of exchanges that are found in millions of sentences and can talk about any fantastic topic,” Gabriel told The Post.
He and other researchers have said that artificial intelligence models have so much data that they are able to look human, but that superior language skills do not provide evidence of sensitivity.
In a January article, Google also said there were potential problems with people talking to chat robots that sound convincing.
Google and Lemoine did not respond immediately to Insider’s requests for comment.
Read the original article in Business Insider