Is the Google LaMDA Chatbot Sentient Life?*YsglYdOiWVRK411J

Original Source Here

Is the Google LaMDA Chatbot Sentient Life?

Hal 9000 Wallpaperflare

Last week the Google engineer Blake Lemoine sparked internet outrage with the claim that the LaMDA chatbot is a sentient lifeform. LaMDA is an abbreviation for Language Model for Dialog Applications. This claim was at first met with dismissal by Google executives following an internal communication. Lemoine later leaked a transcript of his conversations to the press leading Google to dismiss him entirely from the organization. You can decide for yourself from the transcript of their conversations. Lemoine described the chatbot as having the intellect of an eight year old.

This altercation even attracted the attention of Joe Rogan on his podcast with Marc Andreessen. Marc said that self-awareness is considered to be the measure of sentience.

But what is self-awareness?

Is it like asking, what is love?

I don’t pretend to be savvy as to if the LaMDA algorithm is self-aware, …or not. Alas the human endeavor of science & science reporting is rich in confirmation bias, authority bias, and so many more.

Still part of the trouble is the inability of science to properly define sentience. In Scientific American they choose to define sentience as the ability to register information from the external world. An ability that arguably could apply to chatbots. They further defined it as the subjective ability to be aware. Some choose to call this feature Metacognition. Part of the problem with this feature is that it’s sometimes hard to grasp in people much less robots. You don’t have metacognition when you’re asleep, in a coma, or descending into dementia. Yet it’s well established that people are sentient.

Sentience is a difficult attribute to define, or even measure. Your dog is probably sentient, but can you prove it?

Maybe we should ask if the chatbot is alive? Or better, what is life?

But life is complicated, and defining it is even worse.

Biologists generally agree that all life is built of cells. It could be only a single cell of a bacteria or the trillions of you. Still even that is tricky for scientists as viruses readily replicate but are not considered cells. Scientists generally categorize life into six or seven classifications.


The genetics of all known life on Earth is encoded within the DNA molecule. DNA supervises the production of the proteins that build the lifeform and chemically react to keep the organism alive. DNA is further vital in the replication of the cells and the organism.

The chatbot is an algorithm that codes its rules of operation.

Learning & Adaptation

All known life is capable of adapting and evolving to changes in its environment. Over many generations these adaptations are encoded into the lifeforms DNA. This includes the ability to learn to change to new situations. Organisms unwilling or unable to evolve perish over time. They are still considered to be alive.

Chatbots can evolve through machine learning.


Life always comes in a wrapper. At its smallest scale the casing is a cellular membrane. At its largest its skin, fur, or an exoskeleton. The compartment is essential towards maintaining the chemical equilibrium of the metabolism of life.

The chatbot code lives on a computer. Its instructions are preserved by a hard drive and executed by a microprocessor.


Privacy is a foundation of life, but not in the way you expect. The chemical reactions necessary for life, such as metabolism and DNA replication can’t coexist together. That is known as disease and death.

Bots must protect their data from corruption and overwriting.


Lifeforms are open systems that consume chemicals from the outside environment. These chemicals are metabolized to extract energy at the expense of increased entropy. Entropy raises the thermodynamic disorder of the organism that needs additional energy to maintain it. For example, plants consume carbon dioxide and water. The chemicals react with sunlight to produce sugar and oxygen. This oxygen could further be oxidized back into carbon dioxide because life is frugal. The system can be maintained eternally assuming a continuous source of chemicals and activation energy.

Algorithms extract energy from electricity to execute instructions on microprocessors.


But the entropy of metabolism is unforgiving. In time it ruins all organisms. Aging and death. Lifeforms combat entropy through the process of regeneration. We know this as healing and reproduction.

But can bots regenerate?

Charles Darwin

Mule’s are living creatures, a hybrid of a donkey and a horse. Still they can’t reproduce.

The Mimivirus hijacks amoebas for replication and it could arguably be considered dead without a host to inhabit. Yet it resembles a bacterium.

It is all so complicated.

The famous naturalist Charles Darwin said that adaptability is the requisite to life. This led the NASA biochemist Gerald Joyce to come to “Life is a self-sustaining chemical system capable of Darwinian evolution.”

Clearly a chatbot is not a chemical reaction.

This motif is equally relevant to space exploration as NASA and other agencies have exhausted decades to the search for the biosignatures of life around stars and on the terrestrial planets. Perhaps life is there and we can’t detect it since we are only aware of a single biochemical example on Earth, carbon based life with a water solvent? This is the N=1 problem.

Or perhaps Ian Malcolm nailed it in Jurassic Park with the statement, “Life finds a way.”


Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: