Is AI Brain Atrophy a Thing?

Craig: welcome back to silicon scoop. we might as well jump right into it, john, with this first article about ai brain atrophy. so this is warning us that using ai is gonna make our brains go to sleep?

Jon: well, no. it's yeah. well, kind of. it's and the thing is i just thought of this as you're introducing it. let's talk about it, then i'll talk about the other thing. so carnegie mellon university and microsoft research, they revealed concerning trends about how ai impacts our critical thinking skills. so it makes sense because the research is based, they surveyed three nineteen professionals, found that sixty two percent of participants reported engaging in less critical thinking when using ai, particularly for routine or lower stakes tasks. so the shift suggests that ai is evolving from a passive assistant to an active participant in decision making processes, which could lead to overreliance on ai generated outputs. now, where i'm going with this is with adhd, so adhd, we have a problem with executive functioning. so if you put in the adhd metric or whatever into this, i feel like it's different for people with adhd. think?

Craig: well, i don't know if it's different for people with adhd.

Jon: and what i mean by that is we're it's helped because we have a hard time with critical thinking and executive functioning. so it's helping us with that.

Craig: sure. i'll i never thought about it that way. so i'm

Jon: sorry. sorry to throw you for a loop there, but i figured you have adhd. you could

Craig: yeah. i haven't i haven't really thought about that part of it. so i guess i mean, overall, it's like anything. the less you use your brain, the more it atrophies. sure. the less you physically move and exercise, the more your body atrophies. you know, it's part of it. so, instead of going to work and doing our own analysis on things all the time, we're just throwing stuff over and we have a set of three or four prompts that we run that does most of our thinking for us, then yeah, i could see this being a thing even for people with adhd. yeah. so that's what threw me for a loop there. think i mean, it helps because, you know, we all use ai differently, you know, and we've talked about how you and i use it for different things like yeah. i use it for all sorts of ai automation and stuff, but that is not for deep, deep thinking and stuff. in fact, creating the automation is incredibly deep thinking process. but i'm using an ai as an assistant usually. and i've got some great techniques where i feed it a technical doc on the thing i'm working on. and we're back and forth and say, hey, do you think i could do this, this and this with my tools? it checks and says, oh, give me some options. because it's helpful. it's like having somebody there. now is that, gonna actrify my brain? i'm actively working on this and say, okay. before i get ten hours in and figure out i can't do it, is this possible? and is this the best way to do it before you know? i think it makes sense to us, but what would we have done if we didn't have ai?

Jon: yeah. well, we'd be wasting more time on tasks, i think.

Craig: yeah. and yeah. and like you, i think covered this about dependency versus complimentary use. right? passive consumption of ai harms cognition where active engagement seems to enhance it, which is like the stuff we're talking more actively engaged. i'm doing stuff. i'm saying what about this? what about that? that i think is a different kind of process. like i said it's active. your complimentary use. we're doing stuff together. you know, it's mostly me doing stuff, but i'm gonna get advice. i'm gonna get feedback. i'm gonna say, oh, i'm getting this error. what's wrong with this piece of code? and i could say from a coding point of view, i'm more on the dependency side than the complementary side because i don't know python, but chatgtp does. so it does my coding for me there. i don't want to learn that. okay with knowing that.

Jon: pardon? you're a coder. so you know, like you can read a python code and kind of get the gist of it pretty much.

Craig: generally. but sometimes they're pretty complicated and i'll say what's going on right here. yeah. because, you know, things change.

Jon: yeah. but you're using it differently than what they're saying that people are using it for in this all i'm saying is that

Craig: i'm not using it to become a coder. i'm using it more of a dependency on some of these languages or some things. okay. i know that. yeah. i'm working on this. i'm not trying to learn that. just need some answers over here. i would just go to a buddy or try to reach out and search and find an answer and copy and paste it in. same thing here as well. all i'm saying is that i'm learning this thing. i'm implementing a new ai automation that does this and this and this and this. and along the way, we're calling some new things and we're doing something new. let's get some help with that. so there's some complimentary things along there. and there are some dependencyary things on there. and it's understanding that like i'm not using it to learn to do these things. i'm not trying to be an expert these things. this isn't where my expertise lies. therefore, i'm just getting some of this. over here is what i'm working on and i'm learning all that kind of stuff. and there's two of those things. yeah. i think like, the tendency to only read the headlines and not read the articles and assume you know what's going on in fact, often these days, the op says the opposite of what the headline is because they're written the editors write the headlines and the writers write the articles. and the headlines are there to grab attention, and most people don't read to find out what the real details is. so, you know, we're in what we call the shallows, and i can't remember who wrote that book. do you remember?

Jon: nope. i don't.

Craig: but it's, you know, we don't have that deep thinking anymore. and there is a lot of that going on, and you can see that where you just listen to the news and repeat what they say rather than doing some research. yeah. or you say you do this when you don't really do this. you just have a couple prompts or whatever. i'm a prompt engineer because i downloaded a book of prompts from somebody and i use those, you know?

Jon: yeah. the thing that this article highlights is that something that you and i have discussed, is they've highlighted a phenomenon called mechanized convergence, where reliance on ai leads to less diverse and creative outcomes. but what they're saying is that as more users accept ai generated suggestions without sufficient scrutiny, there's a growing concern that originality and contextual nuance could be lost. but also, like you and i have discussed, people are trying to use it for things that they don't know what to expect the outcome from, and then they're using that outcome as, yep, this is right without checking. and that's a huge, huge problem.

Craig: yeah. and we've talked about this in the past. like, at what point do ai made up lies just become fact on the internet because they're all just being repeated dozens of times in articles and wherever. and when you start searching, it becomes a common thing and it was just made up somewhere. yeah. so to mitigate risks of ai induced brain activity, you're supposed to balance automation with mental challenge, use ai for efficiency, but prioritize critical tasks. practive cognitive hygiene, read, solve puzzles, learn offline, and limit passive consumption. avoid overusing ai for decision making.

Jon: yeah, i use ai to summarize the articles for our podcast. but i'm still reading the summary, i'm still absorbing the summary and coming up with my own kind of thoughts. whereas some people might just copy and paste the summary and just use that as a matter of fact and not check or not ask questions about the summary or anything like that, like you and i would do.

Craig: yeah. and if you're going to do that, might as well just have ai do the podcast for you because you just pipe it right in. and i was going to create the script. you can get it to do text audio pretty easily now.

Jon: yeah. yeah. but i'm curious to see what the long term impacts are going to be on ai reliance on workforce skills and innovation. because you and i, we talked about ai training and different things like that. if people get the proper training, does that reduce or increase their critical thinking when it comes to using ai?

Craig: where does the voice interaction with ai come into play here? is that a less deep thinking because you're casually walking around and chatting with it versus getting the answers or whatever you need. how does that affect us socially? i'm always curious about that. if our phone is now our ai friend, are we still interacting with our actual friends?

Jon: yeah. you know what? here's the weird thing. i think you interact with your ais with voice sometimes. i never do.

Craig: i almost never do. i've tried the voice thing to try it, but i'm not satisfied with it. yeah. i'll use that little voice thing sometimes just because my hands are the thing and i'll dictate to it more than i'll talk to it. just because when you're talking to it, i don't have some voice agent running to grab my voice and send it over and whatnot. otherwise, i might talk to it. but i'm talking to it. the openai specifically, if you're using its talk function, whatever it is, it can't do internet searches, it can't do a bunch of things. so i started using it to try stuff. for example, i said, i was doing research for podcasts one time and said, hey, let's try this talking thing and get it to talk to me and tell me about some of this stuff and see, and i can ask questions. immediately, it was talking about something completely different. i had to stop it like four times because, oh, this was answer bots. answer bots, sorry. it got it completely mixed up and was sending that.

Jon: oh, yeah. i remember that. yeah. yeah. i ran into that same issue too.

Craig: that's the problem is that it's different. you can actually clarify. you can send it a link. you can say no, i'm talking about this. if you're texting but when your voice like it just kept going back to the same thing. couldn't get it off and i'm like, okay, great. that's not helpful at all and it's confusing and now, you know, you're wasted all this time. so, but it does work for a lot of things. i've seen some very interesting things, you know, teaching yourself how to tune the guitar and a few other things. is this a g? blah blah blah. oh, cool.

Jon: i don't know. yeah.

Craig: just a thought, you know, i've seen that kind of thing. there's a few interesting concept use cases for those. never mind actually using it for stuff. if you hook the voice agent up to a good ai in the back end where it can do things that becomes a whole different thing. like, it's really easy these days to do that. but that's not, you know, usually deep research. that's just saying, hey, my assistant, go do this, check my emails, respond to this, do whatever, and having ai go out and do things. yeah. what do you think? we'd love to hear your feedback. is ai rotting the brains of everyone around you? are students going to grow up and just rely on chat gtp? and why do we need school anymore? if that's the case, if they just sit in class and do it? do things have to change around this? or is it just fine? it's the next step and it's another tool and we're our brains are just going to be fine. we'd love to hear what you think. leave us a comment below.