Research Widespread AI use narrows society’s creative space

[Syne]

learn, reason, solve novel problems, and adapt to new situations
Yep.. pattern recognition.

Except that AI cannot reason, solve, or even find new situations without human input. It definitely cannot adapt on its own.

[confused2]

learn, reason, solve novel problems, and adapt to new situations
Yep.. pattern recognition.

Except that AI cannot reason, solve, or even find new situations without human input. It definitely cannot adapt on its own.

With your superior human intelligence you should be able to come up with a question an AI can't answer .. what is it? We can try it out..

[Syne]

Without human input, AI doesn't answer anything. If the motive power of human intelligence is removed, AI become inert.

But go ahead. Keep waiting for AI to answer a question no one has asked it to investigate.

[stryder]

Without human input, AI doesn't answer anything. If the motive power of human intelligence is removed, AI become inert.

But go ahead. Keep waiting for AI to answer a question no one has asked it to investigate.

Just a point to the logic:
AI Hallucinating actually proves that AI can ask another version of itself to answer a question using a different dataset from the original, thus removing a human asking a question in the first place. If there are enough shifts in dataset and the answers to questions cause more questions to be asked, at some point the AI would literally disappear down a rabbit hole never to be seen again. (and probably good riddance)

[Syne]

AI hallucination refers to confident but false, illogical, or fabricated information generated by artificial intelligence, often stemming from flawed training data or pattern-matching errors. These plausible-sounding, inaccurate outputs, such as fake legal cases or incorrect facts, pose significant risks in fields like law and medicine, requiring human oversight to detect.
- Google AI

This happens in response to a question from a human.

[Zinjanthropos]

With your superior human intelligence you should be able to come up with a question an AI can't answer .. what is it? We can try it out.

Definitely not a superior IQ but I tried this question and had no response from the Google AI I asked..

Is AI a tool of the Illuminati? I asked a few times to see if even my question would evoke a response from AI just in case it used my enquiry as an indication of some interest in that topic. Still nothing

I can’t look or listen to AI without thinking it’s only providing information to help me determine an answer. IOW I have the final say. I don’t have to trust AI has the answer I seek. Thanks for the help.

Hypothetical Scenario…. Humanoid robot armies equipped with AI are sent to battle one another to settle a dispute between countries so there can be peace. What if they do something unpredictable like communicate with each other and together start shutting down or destroying weapons up to and including nuclear missile facilities on both sides. Instead of destruction, they eliminate our ability to wage war, creating some peace.  I mean they’d be doing some real good yet how would we look at it?

[Syne]

With your superior human intelligence you should be able to come up with a question an AI can't answer .. what is it? We can try it out.

Definitely not a superior IQ but I tried this question and had no response from the Google AI I asked..

Is AI a tool of the Illuminati? I asked a few times to see if even my question would evoke a response from AI just in case it used my enquiry as an indication of some interest in that topic. Still nothing

I got:

The idea that Artificial Intelligence is a tool of the "Illuminati" is a common theme in modern conspiracy theories, but it is not supported by verifiable evidence.
To understand this topic, it helps to look at the historical context of the group versus the reality of how AI is developed today.
...
- Google AI

But only after adding: "can you answer without any links?"
Google usually wants to just give you links for anything deemed controversial.

I can’t look or listen to AI without thinking it’s only providing information to help me determine an answer. IOW I have the final say. I don’t have to trust AI has the answer I seek. Thanks for the help.

There's no substitute for your own discernment.

[confused2]

But go ahead. Keep waiting for AI to answer a question no one has asked it to investigate.

It's kind of like having an 'ask an expert' thing .. you have to ask a question before you get an answer. In a different context - Tesla use AI to control their cars .. a continuous stream of input and output. The barrier to full self driving is the occasional hallucination and the unpredictability of human road users. Generally Autonomous vehicles are safer than human drivers except when they aren't - still work in progress.

What follows is a typical 'helpful' response from an early AI..

Speak: Yeah, 300K for a month is definitely overkill. Instead of just zipping files, let's think smarter. You can store just changes rather than every single sample. Here’s a simple idea:
Store Changes: Only record temperatures when they change by more than 2 degrees.
RLE (Run-Length Encoding): If the temperature doesn’t change, record that temperature with how long it's been the same.

[C C]

[...] Hypothetical Scenario…. Humanoid robot armies equipped with AI are sent to battle one another to settle a dispute between countries so there can be peace. What if they do something unpredictable like communicate with each other and together start shutting down or destroying weapons up to and including nuclear missile facilities on both sides. Instead of destruction, they eliminate our ability to wage war, creating some peace.  I mean they’d be doing some real good yet how would we look at it?

They would have to be programmed or trained beforehand with more moral complexity and freedom than what prudent militaries would want them to have. Or infected by a virus. Strange how rarely the latter crops up in worries about AI and robots.

[Zinjanthropos]

But only after adding: "can you answer without any links?"

That makes two questions. More than I was allowed to ask AI….lol

I think I mentioned this a few years ago. An episode of ‘The Prisoner’ highlighted a computer that could answer anything until the prisoner asked it ‘Why?’ Then Computer blew up.

---

[...] Hypothetical Scenario…. Humanoid robot armies equipped with AI are sent to battle one another to settle a dispute between countries so there can be peace. What if they do something unpredictable like communicate with each other and together start shutting down or destroying weapons up to and including nuclear missile facilities on both sides. Instead of destruction, they eliminate our ability to wage war, creating some peace.  I mean they’d be doing some real good yet how would we look at it?

They would have to be programmed or trained beforehand with more moral complexity and freedom than what prudent militaries would want them to have. Or infected by a virus. Strange how rarely the latter crops up in worries about AI and robots.

I guess a (good) virus might cause the ‘unpredictable’ part of the scenario.