We Compared Eight AI Search Engines. They’re All Bad at Citing News.
https://www.cjr.org/tow_center/we-compared-eight-ai-search-engines-theyre-all-bad-at-citing-news.php
Premium models, such as Perplexity Pro ($20/month) or Grok 3 ($40/month), might be assumed to be more trustworthy than their free counterparts, given their higher cost and purported computational advantages. However, our tests showed that while both answered more prompts correctly than their corresponding free equivalents, they paradoxically also demonstrated higher error rates. This contradiction stems primarily from their tendency to provide definitive, but wrong, answers rather than declining to answer the question directly.
The fundamental concern extends beyond the chatbots’ factual errors to their authoritative conversational tone, which can make it difficult for users to distinguish between accurate and inaccurate information. This unearned confidence presents users with a potentially dangerous illusion of reliability and accuracy.
I did some testing last week and found this out myself related to searches about climate.
AI will only regurgitate prevailing consensus, regardless of whether its authentically based or the truth.
It will use FAKE NEWS for citations.
Re: Re: Re: Re: RFK Jr. Says US Will Know Cause of Autism
By metmike - May 31, 2025, 8:50 p.m.
WARNING:
Only use AI for limited searches and not for critical information that you must be sure is authentic.
AI will only regurgitate the consensus of what is out there. If there is enough bs defining a topic that serves as the consensus, AI will fail miserably for providing independent, objective and authentic information for definitive discernment.
It will fail in other ways too.
AI still operates on the principle of "garbage in.......garbage out"!
+++++++++++++++++
I did this searching using Grok, which is supposed to be the best AI out there:
NASA Confirms Atmospheric Shift Over North America
+++++++++++++
A decent scientific summary by Grok but catastrophically fails with its list of resources/citations/search results to define this topic which includes numerous copies of the Fake News article written by the anti science, climate activist, Ayman Rani!
Experts found evidence that scientific citations in the MAHA report were generated by AI, which experts slammed as ‘sloppy’ and ‘shoddy’
Busting AI's junk climate science:
By metmike - May 31, 2025, 12:20 p.m.
I'm going to repeat the story on the first page that got this thread started because its:
Fake News!
It's an example of how a very skilled person can use the technology of the internet to push whatever propaganda they want. Lot's of impressive graphics and sciencey sounding verbiage along with constantly siting a credible source but NEVER ONCE showing an actual study or link from that source or any other source as evidence of their FAKE or distorted/exaggerated impacts for the Atmospheric shift over North America.
That's because the vast majority of the impacts of their "Atmospheric Shift" are completely manufactured.
By Ayman Rani
AI Overview
NASA studies have indicated shifts in North America's atmospheric patterns due to climate change. These shifts are linked to factors like the movement of ice and groundwater, which are altering the Earth's rotation speed and impacting weather patterns. Specifically, the lengthening of days due to these changes is accelerating, and NASA predicts this trend will continue or even accelerate if greenhouse gas emissions increase.
Earth's Rotation:
NASA-funded studies have shown that changes in ice and groundwater, driven by climate change, are affecting how fast the Earth rotates.
++++++++++++++
If we go to their link/source to find out more about what sounds like a very serious problem, we get this:
"They found that, since 2000, days have been getting longer by about 1.33 milliseconds per 100 years, a faster pace than at any point in the prior century. The cause: the accelerated melting of glaciers and the Antarctic and Greenland ice sheets due to human-caused greenhouse emissions. Their results were published July 15 in Proceedings of the National Academy of Sciences.
“The common thread between the two papers is that climate-related changes on Earth’s surface, whether human-caused or not, are strong drivers of the changes we’re seeing in the planet’s rotation,” said Surendra Adhikari, a co-author of both papers and a geophysicist at NASA’s Jet Propulsion Laboratory in Southern California."
+++++++++++++++++++
"1.33 milliseconds per 100 years"
How much is that really?
There are 1,000 milliseconds in a second. So 1.33 milliseconds every 100 years is 1.3 seconds every 100,000 years.
There are 31,536,000 seconds in a year. They are telling us that in 100,000 years there will be 31,536,001 seconds.
They blow that up into something they claim is part of the climate crisis doom. Even if it were 1,000 times greater than that..........1.3 seconds every 100 years, where is the crisis to our planet?
This came from Google's Artificial Intelligence. Does that sound very intelligent to you??
AI just regurgitates what is out there to the best of its ability, using consensus. This accurately reflects what the majority opinion is. Good indicator much of the time.
However, when the majority are wrong, THEY ARE WRONG, no matter how many there are or how much power they have as message gatekeepers.
++++++++++++++++++++++++++++++++++++++++++++++
Let's add to this:
AI stated above:
metmike: Is that a good summary of the OBJECTIVE/AUTHENTIC science? Or is it a summary of one side of the science which represents the consensus of messages on the internet from the gatekeepers that only tell us BAD THINGS and that this is a CLIMATE CRISIS? AI only knows the latter.
The indisputable, authentic HONEST science is that this is a climate OPTIMUM for almost all life.
+++++++++++++++++++
Let me demonstrate how DIShonest scientists manipulate the science to trick us into believing otherwise because of their cognitive bias that uses the ANTI science method.
It was well agreed on by scientists in every field and proven without question based on observations and research from geologists, paleo-climatologists and other scientific fields that there was a very significant warming between 11,000-5,000 years ago that exceeded the current warm period by a couple of degrees C. in the higher latitudes of the Northern Hemisphere, exactly where the greatest warm temperature anomalies are today. There was clearly LESS Arctic ice then.
Re: Re: Re: Re: Re: April 2025 global temperature
By metmike - May 6, 2025, 5:41 p.m.
Holocene climatic optimum
https://en.wikipedia.org/wiki/Holocene_climatic_optimum
Not only was it warmer then, the beneficial CO2 was at dangerously low levels. The optimal level of CO2 for life is 900 ppm, double this. Calling it a pollutant is a political lie, contradicted by authentic biology/science.
CO2 dropping by 150 ppm instead of increasing beneficially by 150(releasing CO2 that used to be in the atmosphere previously) and none of us would be here right now .The already extremely low CO2 prior to the current climate optimum era for our planet was causing massive crop failures and famines..........
The planet is currently recovering from a biological CO2 famine! Almost all life would prefer much more CO2 and temperatures a bit warmer. Our dishonest consensus science refuses to dial in the profoundly positive response to increasing CO2 from the biosphere, crops and VAST majority of creatures.
https://www.marketforum.com/forum/topic/69258/
At least there is enough legit science on the internet left from before climate science was hijacked and climate history rewritten (by groups like the IPCC) that AI is stable able to regurgitate that accurately. And also from honest climate scientists.
+++++++++++
So this is how dishonest scientists get around that. They use computer models to REWRITE climate history. Throw out the many, many thousands of observations of the empirical and anecdotal data from objective climate science. Throw out the 100+ studies that showed something they don't want to be true and program a model with information that tells it that never happened and call that a "Mystery Solved"
This is an outrage and blatant violation of the scientific method but they not only get away with it, they now represent the consensus in this particular field.
February 16, 2021
https://www.rutgers.edu/news/important-climate-change-mystery-solved-scientists
The long-standing mystery is called the “Holocene temperature conundrum,” with some skeptics contending that climate model predictions of future warming must be wrong. The scientists say their findings will challenge long-held views on the temperature history in the Holocene era, which began about 12,000 years ago.
“Our reconstruction shows that the first half of the Holocene was colder than in industrial times due to the cooling effects of remnant ice sheets from the previous glacial period – contrary to previous reconstructions of global temperatures,” said lead author Samantha Bova, a postdoctoral researcher associate in the lab of co-author Yair Rosenthal, a Distinguished Professor in the Department of Marine and Coastal Sciencesand Department of Earth and Planetary Sciences at Rutgers University–New Brunswick. “The late Holocene warming was indeed caused by the increase in greenhouse gases, as predicted by climate models, and that eliminates any doubts about the key role of carbon dioxide in global warming.”
How temperature evolved during the Last Interglacial and Holocene eras is controversial. Some data suggest that the average annual global temperature during modern times does not exceed the warmth in the Holocene’s early warm period, called the “Holocene thermal maximum,” which was followed by global cooling. Meanwhile, climate models strongly suggest that global temperatures have risen throughout the past 10,000 years.
“The apparent discrepancy between climate models and data has cast doubts among skeptics about the role of greenhouse gases in climate change during the Holocene and possibly in the future,” Rosenthal said. “We found that post-industrial warming has indeed accelerated the long and steady trend of warming throughout the past 10,000 years.
+++++++++++++++++++++++++
They programmed a model to completely wipe out the Holocene climate optimum warmth, and completely erase the Medieval Warm Period and Little Ice Age that happened with 100% certainty just based on historical events that recorded the weather/climate during those times as well as thousands of scientific data/evidence. Instead they have an intentionally programmed model that produced the bold/underlined output/result above. And they called it science. It's phony baloney from very dishonest scientists!
The IPCC, InterGOVERNMENTAL Panel on Climate Change, a political organization was the first to start doing this in the 1990's. Today, their annual reports serve as updated passages for the climate bible for climate religion worshipers.
I documented them doing this and debunked them countless times here:
Science corruption (IPCC)
Started by metmike - March 20, 2023, 5:50 p.m.
https://www.marketforum.com/forum/topic/93926/
++++++++
IPCC ?
Started by 7475 - Feb. 4, 2024, 8:08 p.m.
https://www.marketforum.com/forum/topic/102088/
+++++++++++++++++
This post shows it extremely well:
The myth of affordable green energy is over
By metmike - Oct. 12, 2023, 11:19 a.m.
In this age of Fake News, DISinformation and propaganda, AI is more frequently being used as a very convincing tool to pass it on.
Instead of using an AI based search to debunk or fact check sources and news at what we think is a trusted AI search engine, we sometimes get bad data and the complete opposite. The AI ends up making us even dumber!
“The sudden prominence of AI-generated content in fact checked misinformation claims suggests a rapidly changing landscape,” according to a paper released online this month.
AI images are becoming a big part of the misinformation ecosystem, but real images taken out of context remain a major issue.
+++++++++++++++
https://globalwitness.org/en/campaigns/digital-threats/ai-generated-is-the-new-fake-news/
Generative artificial intelligence tools allow anyone to quickly and easily create massive amounts of fake content.
https://akademie.dw.com/en/generative-ai-is-the-ultimate-disinformation-amplifier/a-68593890
https://priceschool.usc.edu/news/ai-election-disinformation-biden-california-europe/
+++++++++++++++
Sorry but this will not work for most people.
Most people give things that they want to believe in a free pass.........regardless of whether its the truth or not. AI will just make it easier for them to believe and less likely to be skeptical of.
Well, I find Chatgpt enormously helpful. My brother also warned me to be careful about how I use it citing flaws on a subject that he is an expert on.
But I recently went on vacation and was looking for a book to read for pleasure. I went on Chatgpt and stipulated that I wanted non fiction, biographies or history, and only NY Times best sellers list. Presto!! 10 books listed with a short summary of each. 5 of them I had already read. I picked up 2 of them and have enjoyed reading them immensely.
The Biography of Garfield by Candice Millard - The whole book was wonderful. I'll share 2 points that I found interesting.
Garfield did not die from the assassin's bullet. He died from the idiot doctors in the US who probed for the bullet with their bare hands and infected him. We were 20 years behind Europe at the time. The "smart" doctors in the US in 1881 thought it absurd that invisible germs that you cannot see could kill a person. Like many soldiers of the civil war at the time walking around with a bullet in them 16 years after the war, the bullet missed vital organs. He would have recovered completely.
The 2nd thing that I found enlightening was that even though Garfield was an abolitionist, his death was the first moment of healing after the civil war between the north and the south. "Our president has been killed."
The other book I picked up is called "The Wager" - (A tale of Shipwreck, Mutiny and Murder) - The Wager was a British war ship at the outset of a war with Spain. I'm only a little way into it but it is engrossing.
Thank you Chatgpt. I've used it for other practical things but I recognize the caveats.
Thanks very much, joj!
Thats a great point and with an actual example of how this was very useful for you based on your search and allowed you to learn some new things.
My wife uses a paid version of Grok that she thinks is flawless for chemical formulas so I checked it for her and told her to be very careful.
A month ago, I wrote a long letter of recommendation for 1 of my high school chess players to help him get a national scholarship:
https://www.bryancameroneducationfoundation.org/scholarship and she wanted to run She ran it thru AI like she does her documents when I was still not on my final draft.
It missed all sorts of things that I found when I was preparing the final draft and added some bad things.
Its a great technology for many things but has too many flaws right now for me to count on as reliable for important work.
JFK Jr. learned that the hard way!
I really think that some people are enamored with knowing that they are using a new technology and make assumptions based on that.
You‘re a pretty sophisticated and level headed/realistic guy so I’m not referring to you.
Here is another AI usage.
I play a game called Go. It originated in China and is over 2500 years old (even older than chess). AI now crushes the world champion and in fact, strategies and tactics thought to be self evidently sound for centuries, are no longer so. But it is useful even for an amateur like me. I play online for free against other players at my level all over the world. For 4$ per month I have access to an AI feature that allows me to review my games and move by move shows me my mistakes and illustrates why. It's a great educational tool and I'm sure it also is available in chess.
The power of AI is said to be doubling every 7 months !!! If that is true, in 3 years it will be 32 times more powerful than it is today. It is true that garbage in garbage out is something to be weary of. But good input entered in......???
I believe we will see huge breakthroughs in medical research in the coming years, And the economic productivity gains are likely to be great. Perhaps even great enough to overcome having a moron in the White House implementing stupid economic policy.
On the other hand, those individuals with evil intent might bring on a dystopia.
Steven Hawking is alleged to have said something along the lines of: "Just as a dog is aware that its owner is smarter but is unaware of how big the gap is, so too are humans aware that AI is smarter than humans but we are unaware of how large the gap is."
He speculated that AI could end poverty and disease. But he also thought it could be the end of the human race. May you live in interesting times....
joj,
Thanks for participating in what is a fascinating discussion about what you stated well, AI and how it applies to these most interesting times.
I was asked to speak to a group of Boy Scouts from Troop 350 last month and that was exactly my theme.
Kings and Queens didn't have it this good 200 years ago. I'm typing a message on a forum, responding to a friend in NY right now about Artificial Intelligence right now. If this were 1825, what would I be doing to stimulate my brain?
Read a book, of which I had a limited number on a limited number of topics about things that mankind had a limited amount of knowledge/understanding about.
If there's anything that people can understand about me, it's that I love to access new information to learn and to share it. 2025, compared to all previous generations is the golden age for having access to what is likely more than a quadrillion times the information we had at our fingertips(not at a library) that our ancestors had.
And we get the latest news in a flash from countless sources. 200 years ago, in many places they found out a week later about important news that we can know in seconds.
Forecasting the weather is just 1 example of what the galactic advances have done for mankind:
We've actually been using a form of AI in this field for decades. I remember Professor Portman telling us in my first m meteorology class that it would take a human years to solve all the equations for 1 run of the NGM (Nested Grid Model).
And this is the perfect sort of thing we can look up and get a solid AI response to:
how long would it take to solve the equations in 1 weather model for a human
How to make the world a better place (things to be grateful for)
Started by metmike - Jan. 9, 2020, 10:23 p.m.
https://www.marketforum.com/forum/topic/45623/
++++++++++++++++++=
I really think that if people appreciated the enormity of this, it would offset a lot of negatives in their psyche which cause depression and bad vibes dragging them down!
Just typing this post has been an uplifting endeavor for me!