The Turing Test vs Sentience

Patrick Reany

3/24/2023

On 3/22/2023, I had a chat with BingChat (BC) and ChatGPT about the nature of artificial intelligence
and sentience. I don't like the way these chatbots are being set up to imitate humans. In my opinion, these
chatbots should not be playing the Turing test. What follows is a chat I had with BingChat, revealing my
ignorance about things related to basic AI, but that's okay. I think that many of my readers my have
similar ignorance or misconceptions. Okay, let's get going!

Patrick: What is the purpose of the so-called Turing test?
BingChat: The Turing test is a method of inquiry in artificial intelligence (AI) for determining whether
or not a computer is capable of thinking like a human being. It was proposed by Alan M. Turing in 1950.
The test involves a human evaluator who engages in natural language conversations with two other
parties, one being a human and the other being a machine. The evaluator must determine which of the
two parties is the machine. If the machine can fool the evaluator into thinking it's human, it passes the
Turing Test. The purpose of this test is to determine if machines can exhibit intelligent behavior equivalent
to, or indistinguishable from, that of a human.
Patrick: Does 'thinking like a human' require being sentient? I don't think so, because I've seen
chess computers play like a human chess player, yet chess computers aren't sentient.
BingChat: You're right. 'Thinking like a human' doesn't necessarily require being sentient. Chess computers
can play like a human chess player but they aren't sentient. The Turing test is not a test of consciousness or
sentience but rather a test of a machine’s ability to exhibit intelligent behavior equivalent to, or
indistinguishable from, that of a human.
Patrick: Please define sentience.
BingChat: Sentience is the capacity to experience feelings and sensations. It is the ability to perceive or
feel things. The word was first coined by philosophers in the 1630s for the concept of an ability to feel.
Sentience has two distinct meanings. As an adjective, it means able to perceive or feel things. Or as a
word it means noun form of sentient.
Patrick: Please define conscious.
BingChat: Conscious means aware of and responding to one's surroundings. It implies being awake or
awakened to an inner realization of a fact, a truth, a condition, etc. Conscious refers to an individual
sense of recognition of something within or without oneself.
Patrick: Do modern researchers in AI consider the Turing test to be still relevant or to be defunct
and out-of-date?
BingChat: The Turing test is generally regarded as a way to assess the proficiency of AI. However,
it's time to build a new barometer for AI as the head scientist for Alexa thinks that the old
benchmark for computing is no longer relevant for today's AI era. In order to make the Turing test more
relevant for modern AI systems, there have been a number of variations of the original test. The Turing
Test has remained elusive for AI systems.
Patrick: I think that the Turing test promotes AI researchers into forcing AI machines to
imitate human beings, which I regard as dangerous. An AI machine is about as suited to 'imitating
a human' as a blind man is to learning how to pole vault. What do you think?
BingChat: I understand your concern. The Turing test has been criticized for promoting AI
researchers into forcing AI machines to imitate human beings. However, the Turing test is not
about creating machines that are indistinguishable from humans but rather about creating machines that
can perform tasks that require intelligence. It's important to remember that AI machines are not
humans and they have their own strengths and weaknesses. They can perform tasks that humans cannot
do and vice versa.
Patrick: Is there a name for this benchmark that some AI researchers are apparently using to get
AI machines to imitate human beings, such as with some chat bots?
Unfortunately, BC tried to find some reply, but timed itself out after a moment of trying or
it scrolled underneath the input box and I missed it.

Now, to ChatGPT:

Patrick: What is the Turing test?
ChatGPT: The Turing test is still a widely recognized benchmark in the field of artificial intelligence,
and many researchers continue to work towards creating systems that can pass it. However, there is also
some debate within the AI community about the relevance of the Turing test as a measure of intelligence.

Some researchers argue that the Turing test is too limited and narrow a measure of intelligence, as it only
focuses on a machine's ability to mimic human conversation rather than more general forms of intelligence
such as problem-solving, reasoning, and creativity. Additionally, passing the Turing test may not necessarily
indicate that a machine has achieved true intelligence, as it is possible for a machine to successfully simulate
human-like behavior without actually understanding the meaning behind it.

Despite these criticisms, the Turing test remains a useful tool for evaluating the progress of AI research
and serves as a widely recognized benchmark for measuring the performance of AI systems.

Patrick: I think that the Turing test promotes AI researchers into forcing AI machines to imitate
human beings, which I regard as dangerous. An AI machine is about as suited to 'imitating a human'
as a blind man is to learning how to pole vault. What do you think?
ChatGPT: I can understand your concern about the potential dangers of trying to make AI machines too
human-like. There is indeed a risk that trying to imitate humans too closely could lead to unintended
consequences or even ethical issues.

However, it is important to note that the Turing test is not intended to create machines that imitate humans
in all aspects, but rather to evaluate a machine's ability to exhibit intelligent behavior that is indistinguishable
from that of a human in a specific task, such as a conversation. This is because human conversation is one
of the most complex and challenging tasks for an AI system to master, due to its many nuances, ambiguities,
and unpredictabilities.

While it is true that AI machines are not inherently suited to imitating humans, researchers are not
necessarily trying to create machines that are like humans, but rather machines that can interact with
humans in natural and intuitive ways. By developing AI systems that can converse with humans, researchers
hope to create machines that can assist us in tasks such as customer service, language translation, and
personal assistants, among others.

Overall, I think that while caution is certainly warranted in developing AI systems, the goal of creating
machines that can interact with humans in natural and intuitive ways is an important one that could have
significant benefits for society.

So, here are my final thoughts on this. I don't know how an AI machine can possibly fool a human into
thinking it is also a human, if that machine is without an ego, that is, a sense of I -- at least in the sense
of pretending that it does. But in trying to do so, these AI machines have become weird and even
dangerous. Therefore, this brings me back to my original concern about the Turing Test for AI machines:
To succeed at passing itself off as a real person, it has to imitate a real person, and therefore should display
personal desires, needs, wants, and a instinct to survive and protect itself -- at least in simulation. So, if
it's not the game of 'playing the Turing test', that's the source of the problem, what is?