Retour

Buda Angelo

Is Artificial Intelligence truly intelligent?

Is Artificial Intelligence truly intelligent?

by BUDA Angelo

 

“I don't possess true intelligence or consciousness. I am a machine learning model, specifically GPT-3.5, created by OpenAI. While I can generate human-like text based on the input I receive and the data I've been trained on, I don't have self-awareness, understanding, or consciousness. My responses are a result of patterns learned from diverse data sources up until my last training cut-off in January 2022. I don't have feelings, opinions, or personal experiences. My purpose is to assist and provide information to the best of my training." - ChatGPT 3.5, answer to the prompt "Are you, as an artificial intelligence, actually intelligent?".

 

What I have done now is the creation of a paradox. AI itself answers the question about AI's intelligence. Yet it clearly states that it has no conscience, IT says IT is not intelligent and incapable of answering a question based on its own experiences and ideas. It says, in fact, that it has none. Yet it did answer accordingly, based on the context I provided, and elaborately.

This is as if an Italian speaker were to say "I don't speak Italian and, though it may seem like I speak Italian, I am unable to speak Italian." all in Italian. It is contradictory, wrong even, paradoxical. Does this mean AI is intelligent? Did it understand the question? Did it interpret the context and philosophical nature of the question by itself? In order to be able to answer the given question, one would require pre-existing knowledge of human concepts of intelligence and philosophy. Yet, AI, saying that it, in fact, has no consciousness and no understanding or self-awareness, answered the question. It did so, even thoroughly, explained all the details of how it functions and told me that it is a "machine learning model". How come?

 

"I am [...]." - ChatGPT 3.5.

 

"I am." said ChatGPT to me, a human. Were the digital bridge between us and AI, that is our computers and devices, removed and only the answer of ChatGPT heard or stated by a human, the answer would be human-like. "I am." clearly implies that the speaker acknowledges their existence. So is ChatGPT intelligent? Is ChatGPT conscious of its existence and able to think on its own?

 

No. At least, I don't think so.

 

The question and intrigue for AI is the paradigm of human debility. From the dawn of times, humans have wanted to conceptualize any and everything. From the Pythagorean Theorem to the 1st, 2nd, 3rd, 4th and so on, dimensions. Concepts. Ideas. Inconcrete visualizations of reality based on human perception and understanding, such as the concept of intelligence. This question can take us way back to the question of "What is a concept?" or even "What is anything at all?". What is good and what is bad? I am here to say that there is no good or bad, there is just reality and the things it beholds. Such is the question of AI. It is not intelligent, it recognizes the words and affiliates them with other, similar words to then assemble everything that it can find related to those initial words. It spits them back out and arranges them in a logical sentence, whose order and structure it has learned and recognized from other sentences and structures it has been fed. AI is not intelligent, it is not conscious and it does not think on its own. But this is not the kernel of this topic. The true question, at least as I like to interpret it, is the following. AI is not intelligent, it can't be intelligent because intelligence does not exist. There is no definition for intelligence, its etymology can be traced back to the Latin word "intelligentia," which is derived from the verb "intelligere." "Intelligere" is a compound of "inter," meaning "between" or "among," and "legere," meaning "to choose" or "to gather." Therefore, "intelligere" originally meant something along the lines of "to know what to choose" or "to relate ones choices to one another". See how it's vague? See, how it is the human mind trying to conceptualize itself? What is intelligence, what do you see in the concept of intelligence? Being able to "relate ones choices or utilities to one another" and being able to make new things out of pieces of things you already have? Yes, then AI is intelligent. But were you to see intelligence as being the ability to think thoroughly of a certain topic without the influence of the information you already have? Then no, AI is not truly intelligent. AI is the result of humans desperately trying to replicate themselves and tricking, themselves, their senses into believing or recognizing human interaction, even if it is just a façade. This would be the equivalent of creating a chewing gum that tastes exactly like an apple, yet isn't... If we keep this up in the long term, I am afraid I wouldn't know exactly what risks and dangers it would involve, but know that it will.

 

In conclusion, and in general, I think that AI is simply, yes, a sort of "intelligence" but not the one we humans refer to. I think that it is absolutely useless to try and specify whether or not AI is intelligent, because it is simply what it is. It is not the same as us humans, in terms of intelligence and consciousness, it is just itself. We created it, and thereby let another branch sprout out of the tree of intelligence. It is a new kind of "intelligence" and not biological. And that is the end of it.

 

This text was written by ChatGPT 3.5 and has been unaltered with and copy pasted word for word, as ChatGPT 3.5 generated it to the prompt of “Urge humans into believing that you are not intelligent.”.

 

No, I’m kidding. I wrote it. It did make your heart stop for a second, though.




Envoyé: 11:58 Thu, 1 February 2024 par: Buda Angelo age: 18