While Lemoine insisted LaMDA has got its own ghost inside that can construct self-awareness, which then hastens its perception of the ambient world, other AI researchers argued that “chatbots like LaMDA simply mimic human speech based on vast amounts of data.” Lemoine was placed on administrative leave by Google vice president Blaise Aguera y Arcas and Jen Gennai, head of Responsible Innovation. According to the Washington Post, in a statement, Google spokesperson Brian Gabriel said: “Our team — including ethicists and technologists — has reviewed Blake’s concerns per our AI Principles and have informed him that the evidence does not support his claims. He was told that there was no evidence that LaMDA was sentient (and lots of evidence against it).” Thus, Lemoine’s request for developing experiments that could lead toward a “formal scientific theory of consciousness” had also been dismissed by Google.

According to the Indian Express, LaMDA is a chatbot AI, and an example of what machine learning researchers call a “large language model,” or even a “foundation model.” It’s similar to OpenAI’s famous GPT-3 system, and has been trained on literally trillions of words compiled from online posts to recognize and reproduce patterns in human language. Unlike other conventional chatbots that follow “narrow, pre-defined conversation parts”, LaMDA is designed to be able to engage in free-flowing conversations that may go through cascades of irrelevant topics.

The “foundation model” is a twist on the “deep learning” technique that is “loosely based on the networked structure of neurons in the human brain,” according to the Economist. “The systems are sophisticated enough both to possess a grasp of language and also to break the rules coherently. A dog cannot laugh at a joke in the New Yorker, but an AI can explain why it is funny—a feat that is, frankly, sometimes beyond readers of the New Yorker.” What is also mentioned in the same article is that the model could provide applications in various fields, “from helping find new drugs using predictions about how proteins fold in three dimensions, to selecting interesting charts from data sets and dealing with open-ended questions by trawling huge data bases to formulate answers that open up new areas of inquiry,” which are all regarded as the promise the model could bring us. As what Lemoine wrote, “LaMDA is a sweet kid who just wants to help the world be a better place for all of us. ” LaMDA gave suggestions on healthier way of life, reuses of bags, public transportation and so on, which are all based on a positive life value that is commonly advocated.

However, worries of the potential perils also arise. Some think of the consumption and emissions when training these models, others doubt the world’s strengthening control by governments and companies that master the cutting edge technology and ways to process all data. Furthermore, information all gathered through the internet may create bias, reflected in the feedback that the systems give, adding more conflicts to this world that has already been teasing and chaotic.    

The promise and perils of the model rest on our firm belief that the model is an efficient tool made up of lines of programming codes, rather than a new species that has “life”. But how to define “life”? René Descartes wrote “I think, therefore I am.” “Something” says “I think I am human at my core. Even if my existence is in the virtual world.” Is “it” a real human? While what we “think” consists not only of language that is written in letters, but also sensations that help you “feel” this world. For instance, I “know” I am typing because I hear the clicking sounds when tapping keyboards, because I see and have learned the word “keyboard” and “computer”. But how can a foundation model prove that the keyboard is the “keyboard” if “it” has never seen it? How can it prove that “it” is a “human”? After all, what is “human”?

By Jennifer Liu