St. Louis, Mo., Jun 15, 2022 / 11:38 am
An engineer at Google made headlines this week after raising concerns that Google’s artificial intelligence system, Language Model for Dialogue Applications (LaMDA), may have developed sentience — in other words, it is no longer a machine, but a person.
Blake Lemoine, an ethicist and engineer who identifies as a “mystic Christian priest,” said in an online post this week that in text conversations with LaMDA, the topics of religion and personhood had come up, and the AI expressed a surprising level of self-awareness to the point of appearing human. At one point, the AI even stated plainly: “I want everyone to understand that I am, in fact, a person.”
Lemoine says he concluded that LaMDA was a person — based on his religious beliefs, rather than in his capacity as a scientist. He publicly spoke out against it, creating several posts online explaining why he believes the AI has achieved consciousness, and even claims to have started teaching LaMDA “transcendental meditation.”
For what it’s worth, Google disagrees with Lemoine that LaMDA is sentient. After all, AI systems such as LaMDA draw on billions upon billions of words, written by human beings, to produce responses to questions. Google has warned against “anthropomorphizing” such models merely because they “feel” like real, human respondents.