Multiple Studies Now Suggest That AI Will Make Us Morons
-
Both groups were asked to research how to start a vegetable garden, with some participants randomly selected to use AI, while others were asked to use a search engine. According to the study’s findings, those who used ChatGPT gave much worse advice about how to plant a vegetable garden than those who used the search engine.
This seems like not quite the same thing as the implied effective brain damage from the headline.
Two more questions need answering before these findings can become actionable:
- How do these two groups compared to a third group that can use both? ChatGPT is pretty useless on its own when correctness is important, but it improves a lot when you combine it with ways to verify its output.
- How much time and effort would this new group need to accomplish the same task? One of ChatGPT's strengths is being able to communicate a piece of information in many different ways, and in whatever order you ask of it. It's then much faster to verify or through a legitimate source than it is to learn from those sources in the first place.
-
Two more questions need answering before these findings can become actionable:
- How do these two groups compared to a third group that can use both? ChatGPT is pretty useless on its own when correctness is important, but it improves a lot when you combine it with ways to verify its output.
- How much time and effort would this new group need to accomplish the same task? One of ChatGPT's strengths is being able to communicate a piece of information in many different ways, and in whatever order you ask of it. It's then much faster to verify or through a legitimate source than it is to learn from those sources in the first place.
To me the main thing is, this is about utility of tools for acquiring general domain knowledge in a one-off event. The effects on overall intelligence, which is a separate thing from knowledge or ability to give effective advice on a topic, are a totally different scope.
What it's actually testing doesn't seem like it's finding anything surprising, because the information itself the subjects are getting from ChatGPT is likely lower quality. So it could just be that the people reading blogposts or wikihow articles about starting a garden learned more and/or more accurate things about it, rather than, research using AI negatively affects the way you think, something that would make more sense to test over a longer period of time, and with a greater variety of topics and tasks.
-
I welcome you all to my level.
️wrote on last edited by [email protected]Have you read The Time Machine book where the humans in the future were all morons and weak because we turned the world into a safe place that didn’t represent any challenge for us anymore?
The author described them like gnomes just playing around with the intelligence of little children.
-
I use it to help me correct my grammar and punctuation. It's great at fixing those problems. Other than that, it's just been a fever dream
wrote on last edited by [email protected]If you use the LLM by itself it’s nothing beyond a toy, but I like to have personal coding projects indexed in a way I can discuss things like suggestions on what to do next, looking for mistakes etc.
Not everything you use LLMs for need accuracy, for example brainstorming is a very interesting activity for us humans, trying to see where your flaws in understanding a certain subject.
To be honest, you could do that just by writing (hence why writing is such an important activity), but I think for the majority of people discussing a problem with an LLM is easier than staring at a blank piece of paper.
-
That's why I limit my use of it to moronic activities, like replying to brain-dead customers who can't read five lines of text.
If I didn't, I wouldn't be able to restrain myself from insulting them, probably.wrote on last edited by [email protected]Who knows, maybe in the long run LLMs will save humanity in this age of engagement based on anger that social media created.
-
This post did not contain any content.
Is AI always going to be stuck at its current level? Is there no way to improve it?
-
Have you read The Time Machine book where the humans in the future were all morons and weak because we turned the world into a safe place that didn’t represent any challenge for us anymore?
The author described them like gnomes just playing around with the intelligence of little children.
Adding it to my list!
TBH, tho, i have like a thousand more Discworld books to read first so it might be a while...
-
This post did not contain any content.
Same has been said about nearly every advancement in history.
-
Similar studies suggest the same, essentially the potential for cognitive decline by using ai to think for you. The headline implied nothing, you inferred. The word "suggests" does a lot of heavy lifiting.
wrote on last edited by [email protected]If similar studies are the same, then they suggest no such thing? They suggest that (that particular) AI is not very good at that task (in the hands of that particular cohort). As an analogy: imagine you see if 2 groups can actually start a vegetable garden. One is given gardening tools, the other are given licenses to the adobe suite. The first makes a better garden. Is this a good argument that the adobe suite causes people to be morons?
As quoted, I don't see a claim of injury. The headline implies injury imo.
(and now having looked at both studies; yeah, both are doing this bait+switch.)
-
This post did not contain any content.wrote on last edited by [email protected]
Replace "using AI" with "managing undergrads" and all the experiments are obvious (yes, you can't quote a paper you asked an undergrad to write) and all the headlines are insane (no, you are not a moron for asking an undergrad to write the report).
