I am perfectly fine with the notion of a computer intelligence that we should consider alive and worth protecting (as in, destroying it would be murder of some sort). We are not there yet.
In theory, yes. Even if I cannot we know others can (and have).
An AI cannot do that today or for the near future.
It can certainly write new poems. And what do you mean by designing a car or plane? Engineers have been using computers for that for decades, “Start with this shape, and alter it to be more aerodynamic”, or what whatever. Does coming up with the shape of the vehicle count as “designing” it?
Engineers were also doing it without computers.
Computers are a tool. Not intelligence. The person using the tool has intelligence, not the tool itself.
Not sure how poems figure in to this. Any of us can write a poem. Few of us can write a good poem.
Stringing words together is not necessarily a distinction of intelligence. What makes a good poem is subjective and intangible. Maybe a good place for AI since there is no real delineated rules beyond rhyme and meter it must follow.
Just now, I gave ChatGPT the following assignment: “Write a poem in the style of “The Raven” about the spy balloon”. Here’s what I got (within a few seconds):
Once upon a midnight dreary,
As my thoughts were weak and weary,
I heard a distant, faint commotion,
Intruding on my somber notion.
A spy balloon, it did appear,
Floating high without a fear,
Its purpose hidden, yet I knew,
It spied on us, our secrets too.
The night was dark, the air was cold,
But the spy balloon was ever bold,
Sneaking 'cross the borders slyly,
Gathering intel very wily.
With each minute, it came closer,
Threatening our nation’s composure,
But we were vigilant, we were wise,
We wouldn’t let it take us by surprise.
Our defense was ready, our aim was true,
We fired a shot, and down it flew,
Crashing to the ground with a thud,
The spy balloon, it tasted mud.
Never again would it fly,
Its mission thwarted, it would die,
A warning to all who would spy,
We’ll not be beaten, we’ll not be shy.
And so the spy balloon did rest,
Its mission failed, its fate addressed,
A lesson learned, a story told,
Of a spy balloon, forever cold.
I find it difficult to come up with a plausible argument that denies the label “creative” to the generator of such a text on such an assignment.
If our standard is “computers have to do things that humans do to be considered intelligent, but it can’t be anything that humans do without computers”, then we’ve kind of arbitrarily defined away the possibility of computers being intelligent.
If a computer, programmed to develop an aerodynamic shape, comes up with something the human engineer never even suspected, then surely that counts as “creative” and “design”.
Ok, you use computers in your work.
Do you consider them intelligent? Do you consider them a distinct consciousness? Do you think they deserve legal rights? If yes then why? If no, then why not?
Or, are you being argumentative?
I am using the most advanced goalpost recognition software available, and it is unable to track the rapid movement here.
I guess it is if you think AI is sentient or just a really clever toaster.
Sentience is not remotely synonymous with creative intelligence.
Well, what’s the criterion by which we’re defining “intelligence”. You said “creativity”, but apparently what you mean by that doesn’t include writing poetry. You said that “creativity” would be shown by designing an airplane, but a computer that comes up with a novel, unexpected, and effective shape for an airplane doesn’t count. What does count?
Intelligence and consciousness are different. I think computer consciousness is around the corner though.i don’t think either one is a magic quality of the human brain, and I don’t see AI as using a significantly different process than human brains to do what they do.
Good points.
I could write a program that strings random words together. Is that a poem?
I could write a program that strings words together in accordance with language rules. Is that a poem?
I could write a program that considers rhyme and meter. Is that a poem?
More to the point, has the computer ever been creative in any instance above? Poems tend to evoke emotion. Does a computer understand that when writing? Does it knowing that the word “love” has a 58% evocative rating when written means the computer understands a love poem?
And I am pretty sure no one has a “make me a plane” program yet. They use the computer to design parts and whatnot but there is a reason people are still in the process. If Boeing or Airbus could tell a computer to spit out a new design they would have done so long ago.
ISTM you are talking about expert systems. Are those AI? Can your poem writing program also design a plane? A human can.
I think the “creative” requirement isn’t fulfilled if the computer searched a known design space more thoroughly (and efficiently) than a human could, it doesn’t necessarily count as a creative act. In a way, it’s like the “non-obvious” requirement for invention.
I’ll drop this link to a Nature News article on this topic I posted in another AI thread.
Neither does anyone have a “make me a plane” human. Has there ever, in all of history, been a flyable airplane that was designed by a single person?
I’m willing to bet just one of the Wright Brothers could have made their plane.
IMHO the hardest part of that task is figuring out what consciousness is. We’ve been trying to crack that nut for thousands of years and still don’t have a whole lot to show for it. By the time we get it figured out it will probably not be to hard to program AI’s with that quality should we so choose, or possibly we will find that our AI’s already possess it.
You have been a bit unclear on when you are making claims about what AI can do now, and what is possible.
Do you believe that there is anything going on inside human brains that leads to “understanding” or “creativity” that involves anything other than computation?
I believe most things are possible unless clearly against the laws of physics.
I believe someday what could be considered an AI that has “consciousness” will exist (think HAL or Data).
But what we have today is not that. We have expert systems that are clever and mimic human responses in very narrow areas.
The unanswered question in this thread is “are humans intelligent?”. Including in particular “are humans intelligent enough to determine what is intelligent?”. Until we can establish that, any evaluation of what is or is not intelligence is unreliable.