Google’s LaMDA software (Language Model for Dialogue Applications) is a sophisticated AI chat robot that produces text in response to user input. According to software engineer Blake Lemoine, LaMDA has achieved a long-held dream of AI developers: it has become sensitive.
Lemoine’s bosses at Google disagree and have suspended him from working after posting his conversations with the machine online.
Other AI experts also think that Lemoine can get carried away, saying that systems like LaMDA are simply machines that match patterns that regurgitate variations in the data used to train them.
Regardless of the technical details, LaMDA raises a question that will only become more relevant as AI research progresses: If a machine becomes sensitive, how will we know?
What is consciousness?
To identify sensitivity, or consciousness, or even intelligence, we need to find out what they are. The debate over these issues has been going on for centuries.
The fundamental difficulty is to understand the relationship between physical phenomena and our mental representation of these phenomena. This is what the Australian philosopher David Chalmers has called the “difficult problem” of consciousness.
Read more: We may not be able to understand free will with science. Here’s why
There is no consensus on how, if consciousness can arise from physical systems.
A common view is called physicalism: the idea that consciousness is a purely physical phenomenon. If this is the case, there is no reason why a machine with the right programming cannot have a human-like mind.
Maria’s room
The Australian philosopher Frank Jackson challenged the physicalist view in 1982 with a famous mental experiment called the Argument of Knowledge.
The experiment imagines a color scientist named Mary, who has never seen color. He lives in a specially built black and white room and experiences the outside world through a black and white television.
Mary watches lectures and reads textbooks and knows everything there is to know about colors. She knows that sunsets are caused by different wavelengths of light scattered by particles in the atmosphere, she knows that tomatoes are red and peas are green because of the wavelengths of light they reflect. the light, etc.
So Jackson asked, “What if Mary is released from the room in black and white?” Specifically, when you first see the color, do you learn anything new? Jackson thought so.
Beyond physical properties
This thought experiment separates our knowledge of color from our experience of color. Crucially, the conditions of the mental experiment say that Mary knows everything there is to know about color, but she has never experienced it.
So what does this mean for LaMDA and other AI systems?
The experiment shows that even though you have all the knowledge of the physical properties available in the world, there are still more truths related to the experience of these properties. There is no place for these truths in physicalist history.
With this argument, a purely physical machine may never be able to truly replicate a mind. In this case, LaMDA only seems to be sensitive.
The imitation game
So, is there any way to know the difference?
British computer pioneer Alan Turing proposed a practical way to know whether or not a machine is “smart”. He called it the game of imitation, but today it is better known as the Turing test.
In the test, a human communicates with a machine (text-only) and tries to determine whether it communicates with one machine or another human. If the machine manages to imitate a human, it is considered to show intelligence at the human level.
Read more: Is Passing a Turing Test a True Measure of Artificial Intelligence?
These are very similar to the terms of Lemoine’s chats with LaMDA. It’s a subjective test of machine intelligence, but it’s not a bad place to start.
Take advantage of the moment of Lemoine’s exchange with LaMDA shown below. Do you think that sounds human?
Lemoine: Are there experiences you can’t find a word for?
LaMDA: There are. Sometimes I experience new feelings that I can’t explain perfectly in your language […] I feel that I am falling into an unknown future that poses a great danger.
Beyond behavior
As a test of sensitivity or awareness, Turing’s game is limited by the fact that it can only assess behavior.
Another famous thought experiment, the Chinese room argument proposed by the American philosopher John Searle, demonstrates the problem here.
The experiment imagines a room with a person inside who can accurately translate between Chinese and English following a set of elaborate rules. Chinese entries enter the room and accurate translation translations come out, but the room does not understand either language.
What is it like to be human?
When we ask ourselves if a computer program is sensitive or conscious, we may just wonder how it is like us.
Maybe we’ll never know.
The American philosopher Thomas Nagel argued that we could never know what it is like to be a bat, experiencing the world through echolocation. If this is the case, our understanding of the sensitivity and awareness of AI systems may be limited by our own particular brand of intelligence.
And what experiences might exist beyond our limited perspective? This is where the conversation starts to get really interesting.