On LLMs and cognitive offloading
A newly published study [8] reveals how using AI can negatively affect critical thinking:
The findings revealed a significant negative correlation between frequent AI tool usage and critical thinking abilities, mediated by increased cognitive offloading.
Modern theories of consciousness are increasingly suggesting that mind or consciousness can’t be viewed solely as a function of the brain or body. We use tools and start to perceive them almost as extensions of ourselves, shaping how we interact with the world.
For instance, if you write something down in a notebook, you’ve “offloaded” part of your thoughts, knowing you can return to them later. If that isn’t distributing attention and memory between the brain and external “helpers”, I’m not sure what is.
Using a notepad, along with gadgets that have reminders and calendars, clearly shows that many cognitive processes “extend” beyond our “body”. In scientific terms, this is called cognitive offloading.
When asked whether this is good or bad, I’d reply, “?a dépend, mais c’est inévitable”. After all, we wouldn’t have advanced this far if we hadn’t “felt” the spear we threw, or if we hadn’t relied on books and the abacus.
Even so, when calculators first showed up in schools, teachers were worried enough to organise protests, fearing that people would stop learning arithmetic and lose the ability to do calculations “for themselves”. Over the last 40 years, numerous studies have examined this issue, and they largely agree on one point: commoditising a new tool that lets us offload tasks we once did inside ourselves can indeed reduce our ability to perform those tasks independently: without the tools.
One study examining the potential drawbacks of calculators mentioned this (does it remind you of anything?):
领英推荐
3. Sometimes the calculator delivers misleading information; 4. Students may develop undesirable problem-solving behaviour;
We see a similar set of issues in discussions about LLMs: they can hallucinate (to a far greater extent than calculators), but people are increasingly depending on them, which, as the abovementioned study says, risks undermining critical thinking.
However, it appears that cognitive offloading is unavoidable. I also suspect LLMs will become commoditised, meaning their presence is inevitable. As I see it, the only way to get the benefits and avoid the pitfalls is to use them in ways that improve our thinking — by increasing the variety within the conversation.?
AI is here and is to stay, so we need to find better ways to work, study and learn with it.
The easiest way to do so is to stop asking LLMs questions; have them pose questions to you. [9]
References:
Systems Programmer, ANSI C IPTV, Unix/Linux
1 个月Being able NOT to use a calculator is an advantage. At the very least you will see some jaws drop when you do it. :D
Quality Engineering Leader | Educator | Speaker | SAST board member
1 个月I certainly need to check this. It is no surprise to me that genAI is affecting critical thinking in a negative way. It is far too easy to ask a chatbot for answers than research information on your own. People are prone to picking "easy" solutions unfortunately.