Welcome to DU!
The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards.
Join the community:
Create a free account
Support DU (and get rid of ads!):
Become a Star Member
Latest Breaking News
Editorials & Other Articles
General Discussion
The DU Lounge
All Forums
Issue Forums
Culture Forums
Alliance Forums
Region Forums
Support Forums
Help & Search
General Discussion
Showing Original Post only (View all)Something Disturbing Happens When You "Learn" Something With ChatGPT [View all]
https://futurism.com/artificial-intelligence/learning-with-chatgpt-disturbing-snip-
When people rely on large language models to summarize information on a topic for them, they tend to develop shallower knowledge about it compared to learning through a standard Google search, study co-lead author Shiri Melumad, a professor at the Wharton School of the University of Pennsylvania, wrote in an essay for The Conversation about her work.
The findings are based on an analysis of seven studies with more than 10,000 participants. The gist of their experiments went like this: the participants were told to learn about a topic, and were randomly assigned to either only use an AI chatbot like ChatGPT to do their research, or a standard search engine like Google. At the end, the participants were asked to write advice to a friend about what they learned.
A clear pattern emerged. The participants who used AI to do their research wrote shorter advice, with generic tips and less factual information, while the people who used a Google search produced more detailed and thoughtful tips. The pattern held even after controlling for factors like the information the users saw during their research by showing each group the same facts, or what tools they used.
The findings confirmed that, even when holding the facts and platform constant, learning from synthesized LLM responses led to shallower knowledge compared to gathering, interpreting and synthesizing information for oneself via standard web links, Melumad wrote.
-snip-
When people rely on large language models to summarize information on a topic for them, they tend to develop shallower knowledge about it compared to learning through a standard Google search, study co-lead author Shiri Melumad, a professor at the Wharton School of the University of Pennsylvania, wrote in an essay for The Conversation about her work.
The findings are based on an analysis of seven studies with more than 10,000 participants. The gist of their experiments went like this: the participants were told to learn about a topic, and were randomly assigned to either only use an AI chatbot like ChatGPT to do their research, or a standard search engine like Google. At the end, the participants were asked to write advice to a friend about what they learned.
A clear pattern emerged. The participants who used AI to do their research wrote shorter advice, with generic tips and less factual information, while the people who used a Google search produced more detailed and thoughtful tips. The pattern held even after controlling for factors like the information the users saw during their research by showing each group the same facts, or what tools they used.
The findings confirmed that, even when holding the facts and platform constant, learning from synthesized LLM responses led to shallower knowledge compared to gathering, interpreting and synthesizing information for oneself via standard web links, Melumad wrote.
-snip-
This new study was published in the journal PNAS Nexus: https://academic.oup.com/pnasnexus/article/4/10/pgaf316/8303888
The essay one of the study's authors wrote: https://theconversation.com/learning-with-ai-falls-short-compared-to-old-fashioned-web-search-269760
Sadly, despite the study showing that using chatbots for research dumbed users down, the study's authors hope to find ways to make chatbots force their increasingly dumbed-down users to work a bit harder and thus learn a bit more.
The obvious solution is really the only smart one: Step away from the chatbot.
44 replies
= new reply since forum marked as read
Highlight:
NoneDon't highlight anything
5 newestHighlight 5 most recent replies
Something Disturbing Happens When You "Learn" Something With ChatGPT [View all]
highplainsdem
Saturday
OP
Only for those who take it at face value. I immediately expect something to be wrong-- at least partially and
hlthe2b
Saturday
#1
Yes that's the totally sensible approach ... to distrust the information until it has been corroborated
FakeNoose
Saturday
#25
Also understand how low even that bar was. Remember the disdain for "Google University"
RockRaven
Saturday
#2
I've seen so many people posting "I ran this by Chat GPT and this is what it came up with" or
chia
Saturday
#3
That is so cringe-worthy to me. I can't believe how uncritically people have accepted this
Iris
Saturday
#10
I'll bet Cliff Notes at least had 'some' human review. And the notes were revised over the years.
erronis
Saturday
#11
Whatever chatbot(s?) generated that AI-promoting AI slop answer for you makes no reference at all
highplainsdem
Saturday
#9
This is a message board for people to communicate, not a display for AI slop. If it's treated as acceptable
highplainsdem
Sunday
#41
It was a request saying please, not an order. As for your availing yourself of the technology - it's unlikely
highplainsdem
Sunday
#43
I put my use of any technology whatsoever which will help me to oppose TRUMP
Jack Valentino
Monday
#44
Doing a google search is hardly the best way to learn something to begin with...
Wounded Bear
Saturday
#8
Why aren't you writing those reports yourself, if they're supposedly from you? And why do you want
highplainsdem
Saturday
#20
I emphasized writing for yourself because writing is critical to thinking and shouldn't be turned over
highplainsdem
Saturday
#33
OK. It's clear that your reading comprehension is limited, since we - apparently - agree ...
rog
Saturday
#34
You ignored my pointing out that genAI tools are fundamentally unethical because they're trained illegally
highplainsdem
Saturday
#35
If you had spent a bit more time simply googling, you'd have learned quickly for yourself that natural
highplainsdem
Sunday
#37
Sigh. I haven't "singled you out" - I've posted hundreds of threads here on the pitfalls of using genAI,
highplainsdem
Sunday
#39
I like going down the research holes of Wikipedia, pubmed, phys.org, Medscape, etc.
erronis
Saturday
#13
And there are more and more studies showing how harmful AI is, how much it dumbs users down.
highplainsdem
Saturday
#23