The degree to which people revere AI, as if AI were somehow teach us something we didn't feed it first, is becoming bizarre. The name "Artificial Intelligence" is incorrect to the point of being a farce. Nobody knows what intelligence is (look up "the hard problem" with the quotes), so consequently we have no definition for it. Is a gnat "intelligent"? How about a bird that builds a nest that you couldn't? What we call "AI" would actually more accurately be called "learning programs". Such programs are no more or less a human convenience than computers were 50 years ago.
The real concern is that our learning programs will learn exclusively from the Internet, which is vastly polluted with propaganda. Unlike humans, learning programs are unable to learn from the real world, and so their world view is informed ENTIRELY by whatever propaganda humans post online - propaganda that is already skewed to the point of absurdity. One need only query Google, along with Google's most-favored search result (Wikipedia) to understand how distorted the opinions of learning programs will be. EXAMPLE: We have two highly outspoken young congresswomen in the two parties, both of whom go by their initials. Compare the Wikipedia entries for AOC (https://en.wikipedia.org/wiki/Alexandria_Ocasio-Cortez) vs. MTG (https://en.wikipedia.org/wiki/Marjorie_Taylor_Greene) to see how insanely warped the political "trainers" of learning programs are. That, in combination with the average person's trust in such systems, will be our undoing.
Agreed on all counts. Since it is more in the public eye, I'm trying to scientifically consider the accuracy of current AI answers. Stay tuned through all four parts.
Good points. I tweeted it and look forward to your future articles on this. Very important issues ripe for egregious abuses.
The degree to which people revere AI, as if AI were somehow teach us something we didn't feed it first, is becoming bizarre. The name "Artificial Intelligence" is incorrect to the point of being a farce. Nobody knows what intelligence is (look up "the hard problem" with the quotes), so consequently we have no definition for it. Is a gnat "intelligent"? How about a bird that builds a nest that you couldn't? What we call "AI" would actually more accurately be called "learning programs". Such programs are no more or less a human convenience than computers were 50 years ago.
The real concern is that our learning programs will learn exclusively from the Internet, which is vastly polluted with propaganda. Unlike humans, learning programs are unable to learn from the real world, and so their world view is informed ENTIRELY by whatever propaganda humans post online - propaganda that is already skewed to the point of absurdity. One need only query Google, along with Google's most-favored search result (Wikipedia) to understand how distorted the opinions of learning programs will be. EXAMPLE: We have two highly outspoken young congresswomen in the two parties, both of whom go by their initials. Compare the Wikipedia entries for AOC (https://en.wikipedia.org/wiki/Alexandria_Ocasio-Cortez) vs. MTG (https://en.wikipedia.org/wiki/Marjorie_Taylor_Greene) to see how insanely warped the political "trainers" of learning programs are. That, in combination with the average person's trust in such systems, will be our undoing.
Agreed on all counts. Since it is more in the public eye, I'm trying to scientifically consider the accuracy of current AI answers. Stay tuned through all four parts.
Food for thought, from TomsHardWare.com: Google Bard Plagiarized Our Article, Then Apologized When Caught https://tomshardware.com/news/google-bard-plagiarizing-article
I would be interested in its answer to " define the Scientific Method " .
Bob:
I had done that also, and the AI answer was very similar to what I posted here before:<https://c19science.info/Education/Scientific_Method.pdf>.
We now have ultimate tools for plagiarizing . I saw that one of these services is differentiating itself by including the links to the sources .
In anything physical , I stress the need for explicit quantitative relationships , ie: equations .
From my Heartland talk on the basic basics :
https://cosy.com/Science/QuantTrumpsQual.jpg