Today I want to repeat a previous lecture because it touched on the most important aspect of Artificial Intelligence. This name is suggestive, but in my opinion misplaced.
A better name for the computer-generated output would be Artificial data processing. The "output" of a human being we call intelligent. The output of a machine is should be called "processed data" and not the product of intelligence.
The idea that a computer functions like a brain or vice versa that the brain functions like a computer is already an old metaphor and when we use the term Artificial Intelligence, can a computer be intelligent?
The history of philosophy shows us two routes: Cartesian dualism and what we learned in the Materialism project: physicalism, which you can call monism, the mind can be reduced to the working of the physical brain.
This easily leads to the thought, that matter, material things can generate a mind. So, if our physical brain can produce a mind, wouldn't that be possible for a computer too?
But here we run into a few obstacles. When we talk we understand what we say. When I say "I am a heterosexual person" there goes much more on in my mind than just the creation of the arrangement of five words.
The statement may be related to all kinds of ideas and emotions. When a chat program on a computer says the same, does it have the same experience and feeling for all kinds of connotations?
The answer is simple: NO. Artificial Intelligence only needs to arrange a series words, derived from its input. It does not have any understanding of what it is saying.
There is a difference between syntax and semantics. AI follows the proper syntactical rules to compose texts related to the input, for instance, a question about something.
Based on a massive amount of data and statistical procedures it can answer the question without understanding whatever it is talking about. It has no idea of the semantics, that is, understanding the language.
In philosophical terms, this is about intentionality. Intentionality is the property of mental states to form thoughts or imagine something.
When you hear "restaurant", an image or thought arises in your mind. You think of a room with tables and chairs, plates of food, whether Italian or Greek. An image of a hamburger may also arise in your head.
Thoughts and words are about something. Computers lack this characteristic. They only can convert data into signs, and sentences without any intention or thought behind it.
In that respect, a computer is comparable to a thermometer. A thermometer can indicate the temperature perfectly, but that's about it.
A thermometer does not form a thought about the temperature it indicates. When it is twelve degrees he does not think "This will be a chilly day...".
The point here is that we need to get rid of the myth that the relationship between computer programs and hardware is similar to the relationship between mind and brain.
According to physicalism, which we discussed earlier, the mind with all ideas, thoughts, and emotions is a product of the brain. However, a computer program does not originate from the hardware.
The formal symbols of a computer program are meaningless and were developed independently of the hardware. They don't come from it.
It leaves us to ask ChatGPT, if it has a mind :-) I got an honest answer.
ANSWER - As a machine learning model, ChatPT does not have a brain or a mind in the same way that a human does. It does not have the ability to think, perceive, or feel in the same way that a human does.
Instead, ChatGPT is a program that has been trained to generate text based on a given input. It can produce responses that may seem like they are coming from a person, but it does not have the same mental abilities as a human. -END -
So, we have to keep in mind all the time how limited Artificial Intelligence is. It looks like it behaves like a human, but will never be able to answer the question: What do you feel about what you are saying?
Thank you for your attention again....
Main Sources:
MacMillan The Encyclopedia of Philosophy, 2nd edition
TABLE OF CONTENT -----------------------------------------------------------------
1 - 100 Philosophers 9 May 2009 Start of
2 - 25+ Women Philosophers 10 May 2009 this blog
3 - 25 Adventures in Thinking 10 May 2009
4 - Modern Theories of Ethics 29 Oct 2009
5 - The Ideal State 24 Febr 2010 / 234
6 - The Mystery of the Brain 3 Sept 2010 / 266
7 - The Utopia of the Free Market 16 Febr 2012 / 383
8. - The Aftermath of Neo-liberalism 5 Sept 2012 / 413
9. - The Art Not to Be an Egoist 6 Nov 2012 / 426
10 - Non-Western Philosophy 29 May 2013 / 477
11 - Why Science is Right 2 Sept 2014 / 534
12 - A Philosopher looks at Atheism 1 Jan 2015 / 557
13 - EVIL, a philosophical investigation 17 Apr 2015 / 580
14 - Existentialism and Free Will 2 Sept 2015 / 586
15 - Spinoza 2 Sept 2016 / 615
16 - The Meaning of Life 13 Febr 2017 / 637
17 - In Search of my Self 6 Sept 2017 / 670
18 - The 20th Century Revisited 3 Apr 2018 / 706
19 - The Pessimist 11 Jan 2020 / 819
20 - The Optimist 9 Febr 2020 / 824
21 - Awakening from a Neoliberal Dream 8 Oct 2020 / 872
22 - A World Full of Patterns 1 Apr 2021 / 912
23 - The Concept of Freedom 8 Jan 2022 / 965
24 - Materialism 7 Sept 2022 / 1011
25 - Historical Materialism 5 Oct 2023 / 1088
26 - The Bonobo and the Atheist 9 Jan 2024 / 1102
27 - Artificial Intelligence 9 Feb 2024 / 1108
The Discussion
[13:21] .: Beertje :. (beertje.beaumont): Thank you Herman
[13:21] Particle Physicist Bejiita (bejiita.imako): right on the spot
[13:21] Max Chatnoir: There is one thing I don't understand about ChatGPT. Sometimes it goes beyond its sources and tells you something that isn't true. What is happening in that case?
[13:21] .: Beertje :. (beertje.beaumont): it's a large calculator
[13:21] herman Bergson: What I want to emphasize it the enormous limits of AI
[13:21] Particle Physicist Bejiita (bejiita.imako): i heard Nvidias CEO stating that noone need to learn programming anymore cause the AI will do it for us now
[13:22] herman Bergson: That already is the case Bejiita
[13:22] Particle Physicist Bejiita (bejiita.imako): AI might be useful as help but it can only repeat already existing things
[13:22] .: Beertje :. (beertje.beaumont): do we want that Bejiita?
[13:22] Particle Physicist Bejiita (bejiita.imako): and if noone understand how the machine actually works....
[13:22] Max Chatnoir: But sometimes it doesn't stick to existing things....
[13:22] Particle Physicist Bejiita (bejiita.imako): no we dont
[13:23] Max Chatnoir: I kind of get it if I ask it to invent a phylogeny for a dragon.
[13:23] Particle Physicist Bejiita (bejiita.imako): + programming is fun even u can get stuck now and then and then ChatGPT + my ols friend Stackoverflow are of good help
[13:23] herman Bergson: The most important observation is that AI doe NOT understand what is says...
[13:23] Particle Physicist Bejiita (bejiita.imako): but indeed it is just a machine, if we let AI do everything nothing new will ever be created cause the AI can only reurgitate what we have been creating before
[13:24] herman Bergson: It just searches through massiv amounts of data
[13:24] Particle Physicist Bejiita (bejiita.imako): exactly
[13:24] .: Beertje :. (beertje.beaumont): were does that data come from?
[13:24] herman Bergson: It doesn't come close to intelligent thinking.
[13:24] herman Bergson: Everywhere, Beertje....
[13:24] Max Chatnoir: So if it tells me something that isn't true, it has found that information somewhere.
[13:25] .: Beertje :. (beertje.beaumont): yes, but is that data reliable?
[13:25] .: Beertje :. (beertje.beaumont): can we trust it?
[13:25] herman Bergson: from digitalized books and magazines. Google databases, just name it.
[13:25] herman Bergson: no we can't
[13:25] Particle Physicist Bejiita (bejiita.imako): it might look like it creates new images but to do that it have used already existing images and just patched them together, images often stolen from artists without their concent, artists who have then gotten fired and replaced with said AI. quite sad development. AI uses creators work to get rid of them
[13:26] herman Bergson: We can use AI and depending on the kind of program we can trust it.
[13:26] Particle Physicist Bejiita (bejiita.imako): same with game developers ect
[13:26] herman Bergson: For instance , a specialized diagnostic program in a medical environment can be highly accurate in its diagnoses
[13:27] Particle Physicist Bejiita (bejiita.imako): used in the right way it is indeed very useful but when it is used to replace creators so that the CEO of a company can fire the artists and thus cut cost and increase his own wallet...
[13:27] herman Bergson: But still we need a human to check these result
[13:27] Particle Physicist Bejiita (bejiita.imako): indeed, and if noone need to learn programmming any longer
[13:28] .: Beertje :. (beertje.beaumont): that's like the baker testing his own bread
[13:28] herman Bergson: Well, there still have to be people who learn to program...
[13:28] Max Chatnoir: the human brain works pretty well, but does that mean nobody needs to understand neurology?
[13:29] herman Bergson: yes Max
[13:29] Particle Physicist Bejiita (bejiita.imako): i sav one statement that this might be how it is in the star wars universe. Noone anylonger understands the tech cause its just AI and so technology development completley halts at that point cause noone knows how their existing tech actually works and so it is just copied again and again
[13:29] Particle Physicist Bejiita (bejiita.imako): no new development can ever be made at that point
[13:30] Max Chatnoir: Really? Nobody understands how it works?
[13:30] Particle Physicist Bejiita (bejiita.imako): cause noone understands what is under the hood of their current tech
[13:30] Particle Physicist Bejiita (bejiita.imako): no cause its just AI who do all programming and reuse same things over and over
[13:30] Max Chatnoir: Because somebody else put it there? Didn't they make any notes?
[13:31] Particle Physicist Bejiita (bejiita.imako): the star wars universe seems to have become stagnant in tehnology development for that reason he stated
[13:31] herman Bergson: The main thought of today is that AI may look human in its output, but you have to keep in mind that it is far from human, It is jyst data prosessing using human mad algorithms.
[13:31] Particle Physicist Bejiita (bejiita.imako): cause noone understands anything about their current tech
[13:31] Particle Physicist Bejiita (bejiita.imako): exactly
[13:31] Max Chatnoir: Ah, the AI is doing programming. we can't ask it to explain?
[13:32] Particle Physicist Bejiita (bejiita.imako): its a black box that spits out stuff
[13:32] herman Bergson: You mean, asking it for motivation of its specific choices, Max?
[13:32] Max Chatnoir: But the algorithm is not written down somewhere?
[13:33] Particle Physicist Bejiita (bejiita.imako): sure it works somewhat but as NVidias CEO stated that "NOONE NEED TO LEARN OR UNDERSTAND PROGRAMMING any more CAUSE AI WILL NOW DO IT FOR US"
[13:33] herman Bergson: yes it is typed in by a human.
[13:33] Particle Physicist Bejiita (bejiita.imako): That was just a such insane statement
[13:33] herman Bergson: and created by a human
[13:33] Particle Physicist Bejiita (bejiita.imako): ah
[13:33] Particle Physicist Bejiita (bejiita.imako): + i like creating stuff myself also
[13:33] Max Chatnoir: And the human had some reason for writing it as it was.
[13:34] herman Bergson: But you can create a program that itself creates code based on your input
[13:34] Max Chatnoir: And you can see the code?
[13:34] herman Bergson: I have worked with such a tool for database design many years ago
[13:34] herman Bergson: sure
[13:35] herman Bergson: You could load it in a word processor even sometimes.
[13:35] herman Bergson: ab dthen read it
[13:36] Particle Physicist Bejiita (bejiita.imako): actually already LISP was an attempt to create a language that can extend itself so the computer can write its own software. something about data and statements being all and the same, i dont know LISP so good i can get what that really means however
[13:36] herman Bergson: For instance, I can tell such a tool: input is name address city, generate a notecard with the input.
[13:36] Particle Physicist Bejiita (bejiita.imako): but it was created in the 50s
[13:37] Particle Physicist Bejiita (bejiita.imako): i think
[13:37] Max Chatnoir: OK so what do you think is happening when Chat GPT tells you stuff that isn't true?
[13:37] herman Bergson: Then the tool creats the code and saves me a lut of typing
[13:37] Particle Physicist Bejiita (bejiita.imako): ah
[13:37] Particle Physicist Bejiita (bejiita.imako): yes things like that
[13:37] herman Bergson: It simply makes errors in its data processing
[13:38] Particle Physicist Bejiita (bejiita.imako): I have tried to use it with Unity C# and UE5 C++ a bit, it works but the output is VERY generic
[13:38] herman Bergson: which must be caused by mistakes in the algorithms it uses, which are created by human programmers
[13:38] Particle Physicist Bejiita (bejiita.imako): yes
[13:38] Particle Physicist Bejiita (bejiita.imako): might be
[13:38] Particle Physicist Bejiita (bejiita.imako): shit in = shit out
[13:38] Particle Physicist Bejiita (bejiita.imako): thats how computers work
[13:38] Max Chatnoir: So it's not lying, it's just malfunctioning.
[13:39] herman Bergson: Thaare you have a point Max....
[13:39] Particle Physicist Bejiita (bejiita.imako): a computer just blindly follows instructions with no understanding what it is actually doing
[13:39] herman Bergson: Lying is an intentional act. It is something a human can do.
[13:39] herman Bergson: A computer program just executes its code
[13:40] herman Bergson: There is no intention at all behind it. It even doesn't understand what it is talking about
[13:40] Particle Physicist Bejiita (bejiita.imako): as i heard a statement about CNC machines. The machine have no idea it is a milling machine and it would just attempt to drill right through itself if u give it the wrong instruction
[13:41] Particle Physicist Bejiita (bejiita.imako): search for CNC crashes on youtube, its hilarious and sometimes also dangerous
[13:41] Max Chatnoir: So it is programmed to apologize if you say "that's wrong"
[13:41] Particle Physicist Bejiita (bejiita.imako): + often VEEERY expensive like wrecking a machine for millions of dollars
[13:41] Max Chatnoir: Does it remember that it's wrong?
[13:41] herman Bergson: Besides...lying is a concept that supposes a relation between two people....
[13:42] herman Bergson: No, but some chat sessions ask what you think about the answer. Then you can say...it was wrong. Then it learns
[13:42] herman Bergson: But it is YOU teaching the machine:-)
[13:43] Max Chatnoir: I should ask it again about sex chromosomes in marsupials and see what it says.
[13:43] Particle Physicist Bejiita (bejiita.imako): exacty
[13:44] herman Bergson: A machine can only discover that it is "wrong" when you supply new and extra data that contradict the primary answer of the machine
[13:44] Max Chatnoir: And that become part of its database.
[13:44] Particle Physicist Bejiita (bejiita.imako): ah
[13:44] herman Bergson: Well, if it has insufficient data to process you might get a nonsense answer
[13:45] Particle Physicist Bejiita (bejiita.imako): indeed, i see this with some audio software like UDIO which works but suddenyl can just start blubber
[13:47] herman Bergson: there are two groups in AI, one group believes in AI application for specific subjects ( and we have examples of that)
[13:47] herman Bergson: and there is a group that sti believes in AGI, Arificial General Intelligence.
[13:48] herman Bergson: They believe that AI can really work like the human brain.
[13:48] Particle Physicist Bejiita (bejiita.imako): ok
[13:48] Max Chatnoir: Would we know if it did??
[13:48] herman Bergson: That last group is an end for me.
[13:49] .: Beertje :. (beertje.beaumont): Herman !!!
[13:49] herman Bergson: I wouldn't worry about that Max.
[13:49] herman Bergson: Sorry :-)
[13:50] .: Beertje :. (beertje.beaumont): your av fainted
[13:50] herman Bergson: oh...is because I used the word
[13:50] herman Bergson: lol
[13:50] .: Beertje :. (beertje.beaumont): gaat ie weer.....
[13:50] Particle Physicist Bejiita (bejiita.imako): ㋡
[13:50] herman Bergson: some silly attachment I forgot about
[13:51] .: Beertje :. (beertje.beaumont): looks scary
[13:51] herman Bergson: Better here than in RL, Beertje :-)
[13:51] .: Beertje :. (beertje.beaumont): ja:)
[13:51] herman Bergson: But don't worry. In RL I really feel fine.
[13:52] Particle Physicist Bejiita (bejiita.imako): thats great
[13:52] Particle Physicist Bejiita (bejiita.imako): ㋡
[13:52] Max Chatnoir: good
[13:52] .: Beertje :. (beertje.beaumont): :))) we are glad that you are fine again
[13:52] herman Bergson: Yes I am doing OK :-)
[13:52] herman Bergson: Happy about it myself too :-)
[13:52] .: Beertje :. (beertje.beaumont): :))
[13:53] Particle Physicist Bejiita (bejiita.imako): YAY! (yay!)
[13:53] herman Bergson: OK...that was it for today :-)
[13:53] herman Bergson: Putting AI in its place :-)
[13:53] Max Chatnoir: good one, Herman!
[13:53] .: Beertje :. (beertje.beaumont): thank you Herman
[13:54] Particle Physicist Bejiita (bejiita.imako): well i need to make some more dialogue in Elevenlabs for my Chernobyl UE5 game
[13:54] Particle Physicist Bejiita (bejiita.imako): but first some cards?
[13:54] .: Beertje :. (beertje.beaumont): yes ok
[13:54] herman Bergson: ok
No comments:
Post a Comment