The mind of AI

[Dontaskme]

If you're saying that there is no such "thing" as consciousness, then we diverge there.

There is no such ''thing'' as consciousness.

Things are objects. Objects are not conscious.

[Gary Childress]

If you're saying that there is no such "thing" as consciousness, then we diverge there.

There is no such ''thing'' as consciousness.

Things are objects. Objects are not conscious.

I agree to whatever extent, however, there is "consciousness". It's "something" we talk about. It's "something" that living beings have (especially humans) but that rocks and inanimate matter presumably don't.

[Dontaskme]

I agree to whatever extent, however, there is "consciousness". It's "something" we talk about. It's "something" that living beings have (especially humans) but that rocks and inanimate presumably matter don't.

There is consciousness, but it has no image.

There is much talk of this imageless something.

[zoey]

A year back, the max AI I knew was essay checker and spell checks which had commonly crept into daily internet life. But generative AI and ChatGPT has just been unprecedented. Who thought you could give prompts and have an entire essay and short story written. However, the usage of AI should be raising huge moral and ethical dilemma for the wastage of our most precious resource - water.
"According to a recent report by the Associated Press, AI systems like ChatGPT cause data centers to consume about 500 milliliters of water each time a user poses 5 to 50 prompts or questions. Considering that the chatbot has been the fastest-growing tech site visited on earth since the creation of the internet, that's not an insignificant amount of water, prompting experts and environmentalists to ring alarm bells.
ChatGPT's trained its GPT-4 models on data centers located near corn fields west of Des Moines, Iowa, the report notes, where water from the Raccoon and Des Moines rivers is used. “It (GPT-4) was made by these extraordinary engineers in California, but it was really made in Iowa,” Microsoft’s president, Brad Smith said according to AP.
Since then, the AI revolution has required more water to run its power-hungry hardware —both to power them up and to cool them down. According to official reports, Microsoft's water consumption spiked 34 percent from 2021 to 2022, likely driven by AI-related computing demands. Google, its main competitor, reported a 20% growth in water consumption during the same period."
According to the US Drought Monitor, nearly all Iowa faces some dryness or drought. do we really need such technology?
https://decrypt.co/155965/artificial-in ... ta-centers.
https://www.iowadnr.gov/About-DNR/DNR-N ... %20drought.

[LuckyR]

The main current problem with AI, is that it "learns" from the totality of written material. Which was generated in a biased environment. Thus AI will (and has already) "learn" that the optimal goal is a perpetuation of the same biased scenarios that it "learned" from. Resulting in having AI screen résumés for HR Depts which favor candidates similar to the current workforce. So inanimate, innately unbiased minds will produce learned biased racist results.

[nemos]

It's convenient to judge AI as long as it serves our interests, but what if it ever shows its own interests ... ?

[LuckyR]

It's convenient to judge AI as long as it serves our interests, but what if it ever shows its own interests ... ?

It will depend on whether the program is told that the data it "learns" from is to be followed, such that the data guides it's priorities, or it is preloaded with priorities such that it prioritizes the value of the data it consumes.

Garbage in, garbage out.

[nemos]

As long as AI has no interests of its own, or is programmed to represent someone else's interests, it is just a tool and nothing more. However, many people really, really want to see an AI that has its own personality. Who can argue that a wish will not produce the desired ?

[Iwannaplato]

It's convenient to judge AI as long as it serves our interests, but what if it ever shows its own interests ... ?

It will depend on whether the program is told that the data it "learns" from is to be followed, such that the data guides it's priorities, or it is preloaded with priorities such that it prioritizes the value of the data it consumes.

Garbage in, garbage out.

Or they just create something that learns and who knows where it goes.

[LuckyR]

It's convenient to judge AI as long as it serves our interests, but what if it ever shows its own interests ... ?

It will depend on whether the program is told that the data it "learns" from is to be followed, such that the data guides it's priorities, or it is preloaded with priorities such that it prioritizes the value of the data it consumes.

Garbage in, garbage out.

Or they just create something that learns and who knows where it goes.

Well if it's learning in it's own environment (like the optimal way to win at the old video game Breakout), then it can learn the "best" way to accomplish it's goal. If it's learning from data in our environment, it'll conclude the best incoming freshman for the next Harvard class is a rich, white, lacrosse playing legacy named Trevor.

[Iwannaplato]

It will depend on whether the program is told that the data it "learns" from is to be followed, such that the data guides it's priorities, or it is preloaded with priorities such that it prioritizes the value of the data it consumes.

Garbage in, garbage out.

Or they just create something that learns and who knows where it goes.

Well if it's learning in it's own environment (like the optimal way to win at the old video game Breakout), then it can learn the "best" way to accomplish it's goal. If it's learning from data in our environment, it'll conclude the best incoming freshman for the next Harvard class is a rich, white, lacrosse playing legacy named Trevor.

Unless it decides that the best incoming freshman to be converted into petrie-dish filler is a heavier set Latino who loves darts. It could end up deciding some priorities we just didn't expect.

[LuckyR]

Or they just create something that learns and who knows where it goes.

Well if it's learning in it's own environment (like the optimal way to win at the old video game Breakout), then it can learn the "best" way to accomplish it's goal. If it's learning from data in our environment, it'll conclude the best incoming freshman for the next Harvard class is a rich, white, lacrosse playing legacy named Trevor.

Unless it decides that the best incoming freshman to be converted into petrie-dish filler is a heavier set Latino who loves darts. It could end up deciding some priorities we just didn't expect.

Sure it COULD, but since AI is already used to sort résumés, we already know AI weights future employees into copies of current employees. So if you've had human bias in hiring in the past you'll get future "pass through" bias from an inherently unbiased machine in the future.

[Iwannaplato]

Well if it's learning in it's own environment (like the optimal way to win at the old video game Breakout), then it can learn the "best" way to accomplish it's goal. If it's learning from data in our environment, it'll conclude the best incoming freshman for the next Harvard class is a rich, white, lacrosse playing legacy named Trevor.

Unless it decides that the best incoming freshman to be converted into petrie-dish filler is a heavier set Latino who loves darts. It could end up deciding some priorities we just didn't expect.

Sure it COULD, but since AI is already used to sort résumés, we already know AI weights future employees into copies of current employees. So if you've had human bias in hiring in the past you'll get future "pass through" bias from an inherently unbiased machine in the future.

I guess I'm thinking of more generalized intelligence AIs. One's that are intended to roam more freely or take on a broad range of more opened ended issues and problems. Sorting CVs sounds like a task with a family relationship to chess playing programs. Ask them broad questions, questions from other fields, or even, hey, AI, read the newspapers, pick an issue that interests you and get back to us with your suggestions for solutions,
and these more focused intelligences or 'intelligences' can't even begin. I think they are trying to let some AIs sort of grow like brains, without programming so much as the ability to learn. I'm way to ignorant about this all, but my impression is that they are really shooting at creating minds not task doers and this opens up, well good and bad outcomes from my perspective.

[Philosphicalous]

He got off lightly compared to poor Jacinda Ardern



fce54cec-570a-4508-8cb7-648883b80bd0.jpg

[Philosphicalous]

He got off lightly compared to poor Jacinda Ardern



fce54cec-570a-4508-8cb7-648883b80bd0.jpg

Missed the lump in the crotch