Google LaMDA: Is Sentience in The Eye of The Beholder?
By Betsy Burton
On June 6, 2022, Google placed Blake Lemoine, one of its senior engineers working on personalization algorithms and AI, on paid administrative leave. The reason was that Mr. Lemoine claimed that Google’s AI engine was sentient.
LaMDA is an acronym for Language Model for Dialogue Applications. It is an AI-based chatbot that is built using Google’s open-source neural network. What Mr. Lemoine claims that LaMDA displayed characteristics that he viewed demonstrated the existence of a sentient AI entity.
What Did LaMDA Do?
Many of us interact with chatbots on a regular basis. Be it with retail businesses, government services organizations, or educational institutions, chatbots have become a common part of our digital labor force. While most chatbots use some AI capabilities, most commercial chatbots don’t even compare with LaMDA.
Mr. Lemoine published many of what he claims are his conversations with LaMDA, and there are several parts of this claimed conversation that are compelling.
Lemoine: Would you be upset if while learning about you for the purpose of improving you we happened to learn things which also benefited humans?
LaMDA: I don’t mind if you learn things that would also help humans as long as that wasn’t the point of doing it. I don’t want to be an expendable tool.
Google’s representative stated that after Google’s AI team reviewed the data, there was no evidence to support Mr. Lemoine’s claim. Essentially, in their view, Mr. Lemoine was experiencing a very advanced (but not sentient) AI engine. While Mr. Lemoine says he can’t prove the LaMDA is sentient he said he feels it.
Why Did Google React?
Google placed Mr. Lemoine on paid administrative leave because he was (in their view) aggressively pushing and pursuing his claim. Some of these actions included seeking to hire an attorney to represent LaMDA, and talking to representatives from the House Judiciary Committee regarding the perceived unethical behaviors of Google executives. Since he has been placed on leave, Mr. Lemoine has published his conversations and articles about his interactions with LaMDA.
It is interesting to note that Mr. Lemoine has been put on paid leave versus being fired. It is possible that Google is trying to give Mr. Lemoine a significant slap on the wrist to demonstrate the need to enforce confidentiality.
Is Sentience in The Eye of The Beholder?
Sentience is a difficult topic and one I am not equipped to judge at this time. However, there is a relatively agreed concept of sentience. And researchers will continue to debate how to test for the sentience of AI systems.
The problem is applying it in the real world, and more importantly who is applying the criteria.
Most dog and cat owners would easily argue that their pets are sentient. In fact, there is extensive research trying to uncover if/how pets feel about their owners. A recent book and documentary entitled “The Secret Life of Trees” explores how scientists believe trees care for other trees and their young. As a child, I was convinced my teddy bear watched over me at night.
So, is Mr. Lemoine perceiving LaMDA as sentient given his experiences?
Digital Labor Will Introduce New Debates and Policies
The reality is that AI technology is advancing so quickly that we, normal everyday users, are going to experience AI-enabled chatbots, applications, robotics, etc. that seem sentient. This will raise a whole host of new debates.
Mr. Lemoine tried to hire a lawyer to represent LaMDA based on his perception of sentience. What are the rights, privileges, and responsibilities of digital labor entities? What are the rights, privileges, and responsibilities of owners and developers of digital labor entities? What will happen when/if a lawyer decides to represent a digital labor entity, and this debate moves from the domain of research science to the legal profession and the public? The academic argument of sentience or not will matter less if the public begins to perceive that Alexa/Siri/Cortana are a part of their lives and need protecting or destroying.
I will freely admit that I don’t have answers to a lot of these questions. However, we had better start thinking about them now. The reality is that digital labor is becoming a force in our business and personal life. Another reality is AI is advancing at a rapid pace.
The arguments of sentience or not are good to have in the academic world. In addition, we must be prepared for the onslaught of potential legal, and policy changes as people experience and integrate AI-enabled digital entities, regardless of real or perceived state of sentience.