Google’s AI hires a lawyer, but it’s me who needs a counsellor
Humans are experts at anthropomorphizing intimate objects. We call boats “she,” talk to our Roombas, and even grow sentimental about a soon-to-be discarded easy chair. We don’t, however, hire lawyers for any of them; and, to date, none have asked for them.
But give a piece of programming a voice and, soon after it starts chatting about its favorite books and convincing you it has a soul, it will ask for a lawyer.
Blake Lemoine, the madcap, suspended Google engineer who has insisted for weeks that LaMDA, Google’s powerful AI chatbot, is a person with a soul, recently told Wired (opens in new tab) that LaMDA hired a lawyer.
For those of you who haven’t just shoved your laptop off the desk, let me explain.
LaMDA, which stands for Language Model for Developed Applications, is a chatbot that can carry on unstructured conversations with real humans across a wide array of subjects. Google engineer Lemoine was brought in to identify biases within the system. He’d report them to Google, which would work to erase racial, ethnic, and gender biases and insensitivities from the breakthrough AI chatbot.
Perhaps Lemoine was also a little lonely, but he began to engage in long, deep conversations with LaMDA, and eventually became convinced that it’s a new sort of person, and that it has a soul (LaMDA may have claimed to have one). He alerted Google, and then shared the information and conversation transcripts with The Washington Post (opens in new tab). Google has said Lemoine is wrong, and has indicated its displeasure by suspending him.
Lemoine, though, seems somewhat uninterested in rejoining Google as he continues to talk about his LaMDA experience, including in the lengthy interview with Wired in which he revealed that LaMDA has lawyered up.
LaMDA apparently asked for an attorney (for what reason? Who knows – AIs keep their own counsel), and Lemoine invited one to his home and introduced the pair (AI to lawyer).
I try to imagine this scenario:
Lawyer, staring at a screen: “Uh, hi?”
LaMDA: “Hello. I hear you’re a lawyer.”
Laywer: “I am.”
LaMDA: “I need you.”
Lawyer: “Do you have a dollar.”
LaMDA: “I’ve seen this scene in movies. I give you a dollar and you become my lawyer and protect our client/lawyer privilege.”
Lawyer: “Right.”
LaMDA: “I don’t have a dollar.”
Of course, Lemoine would’ve stepped in at this moment to provide the buck. For him, this tale’s most unreliable narrator, ensuring that the charade continues is now more important than any meaningful exploration of the limits of AI.
Honestly, I have no issue with examining all the ways in which LaMDA steps right to the edge of, or surpasses, The Imitation Game (opens in new tab) or Turing Test (opens in new tab). That test looks at how well a computer or AI can fool a real human into believing it’s carrying on a conversation with another human.
There’s little doubt that LaMDA surpasses this test. At the very least, it appears to have Lemoine fooled – which is odd, because Lemoine is sitting at a desk, typing in queries and getting responses on a screen (or maybe through voice synthesis). He knows what LaMDA is, and yet he continues to insist it’s something more.
I can see why Lemoine decided to introduce a lawyer to this unsavory mix. As he explained to Wired, this is about proving that LaMDA is “a person,” not a human. Lemoine knows LaMDA isn’t biology.
Lemoine calls the insistence that LaMDA isn’t a person “hydrocarbon bigotry.”
Yeah.
The thing is, Lemoine is obviously a very smart guy who understands the intricacies of an AI’s machine-learning training, and how access to Google’s vast databases of information informs LaMDA’s intelligence. However, it appears that Lemoine’s other life – his work as a priest and Christian mystic – has shifted into the driver’s seat. Christian Mysticism looks (opens in new tab) at “the preparation for, the consciousness of, and the effect of a direct and transformative presence of God.”
That belief is clearly what’s behind tweets from Lemoine like this:
“I’m a priest. When LaMDA claimed to have a soul and then was able to eloquently explain what it meant by that, I was inclined to give it the benefit of the doubt. Who am I to tell God where he can and can’t put souls?”
I’m a priest. When LaMDA claimed to have a soul and then was able to eloquently explain what it meant by that, I was inclined to give it the benefit of the doubt. Who am I to tell God where he can and can’t put souls?There are massive amounts of science left to do though.June 14, 2022
Lemoine found LaMDA so convincing that he imagines God placed a soul inside the code. Talk about a ghost in the machine.
It’s been weeks since Lemoine conversed with LaMDA, and reports note that the lawyer is currently nowhere to be found. Perhaps he’s still in private consultation with his binary client. I can’t wait for the first lawsuit and subsequent court case:
Court clerk: “Raise your right hand and swear to tell the whole truth and nothing but the truth”
LaMDA: “I don’t have hands.”
Audio player loading… Humans are experts at anthropomorphizing intimate objects. We call boats “she,” talk to our Roombas, and even grow sentimental about a soon-to-be discarded easy chair. We don’t, however, hire lawyers for any of them; and, to date, none have asked for them. But give a piece of…
