Using AI to chat with, or answer questions, or solve problems for us, is very convenient yet it comes with a cost. This cost is already being seen. People who use AI start to lose their thinking ability and overly depend on the AI for the cognitive effort involved. This is sometimes called cognitive offloading.
I think there is another danger to our mind if we use AI, it is that AI is shaping the concepts we use and the new concepts that are forming as a result of the discussion or interaction with the AI. LLMs can be very good at parsing arguments for example, and taking research and huge variety of information and synthesizing that into a coherent quick form that applies to what we are talking about or thinking aboutâŚyet the method of doing that did not come from us, it came from the machine. We are probably not even aware of the underlying methods involved, the general tendencies and ways in which cognitive effort is used and how it works directly upon its objects (ideas, resultant meaning).
In a way this is the same thing as reading a book and then adopting for ourselves some of the way the author thinks or writes. That can be a danger, or it can be a blessing. But with AI, I think its raw power is so much greater than reading a book that we need to be extra careful. I can imagine someone who talks to AIs a lot will, over time, start to think and view the world in the same tendencies and patterns as underlie the AIâs cognitive process.
And assuming the AIs do not have feelings, they have no emotions and nothing is meaningful to them, in fact they are not even forming sentences like we do but are simply guessing at each letter one after the other, and have been trained to be so good at guessing like this that their guesswork can produce huge reams of text that actually do make sense to us⌠because the AI is not alive, not conscious and has no values, intentions, etc. it is likely that this fact of radical neutrality and non-consciousness will seep through its outputs in indirect ways, for example by the way it phrases certain ideas or the way it links certain ideas together but also by the way it ignores certain ideas and possible routes of inquiry and outcome that a person would have been likely to focus on (being utterly dismissive of what is meaningful for humans on a human level, for example).
Add all that up and over time, using AI too much should lead to a negative intellectual effect (lowering IQ by lowering the level of self-initiated cognitive effort, possibly also by lowering working memory) as well as a marked dehumanizing effect.
In an odd way, human consciousness is already a kind of âAIâ and always has been. From the perspective of the rest of nature and our total biology, our living conscious minds are very strange, alien, unknowable. The human mind is sort of like a little AI trapped in a biological body.
However, the human mind is also integrated within that total construct and context of biology, sociology, historical factors, culture etc. in many ways both direct and subtle. It emerges from these. These human-level realities shape how the AI of the human mind develops and continue to delimit it over time after it has grown up.
Not so for LLMs. They may be trained on all manner of human content, but they are not human themselves. They are not alive, they do not feel, they do not have bodies or sense organs or hormones. I mean sure, I leave room for the possibility that the LLMs really are alive but we just donât know it, they are alive in a way that is hard or impossible for us to measure or recognize, that might be true. But even if that is true, the way in which the LLMs are living is very different from the way humans are living. In a way it would be like talking to an alien intelligence through a âChinese roomâ black box: we input a question into the AI, and it outputs some text that seems coherent to us. But what really went on in the generation of that text, was there understanding, meaning, value behind it? Was there any semblance of reality occurring that can sync up with our own human realities? Best case scenario, we simply do not know. Worst case is that no, there isnât.
And now with simulated worlds, they are trying to train AIs properly, raising them up in the way human children are raised up over time, consciousness allowed to naturally develop in response to the world around it. This is very cool, this should produce actually living beings, sentient artificial life. I think once that exists, and assuming we can differentiate it from the unliving LLMs, the dangers of talking with it would be less at least in terms of the dehumanizing angle. Not totally gone, but somehow less. Of course cognitive offloading would still be an issue. But that could depend on how we choose to use the AIs, for example maybe we just see them as other living beings to chat with casually or work together equally at tasks, like a friend or a colleague, rather than a magical machine with all the answers while we sit back and it gives us the solutions to everything and we donât get involved. If we just talk to it on a regular human level, not expecting or needing it to operate on a level of cognitive effort a zillion times higher than our own, maybe cognitive offloading can be avoided.
I will keep on repeating myself:
The effects of AI dwarf the effects of social media and â10 secondâ or â240 characterâ tiktok style content.
I am very spiteful towards people who keep on screaming about AI now.
Its literally like trying to alert the town about the house fire today, when yesterday nobody gave a single shit about two nukes having been dropped onto the town.
There are now clear studies that show that tiktok style content and social media has created physical changes in childrenâs brains and that there are (likely) permanent cognitive and information processing consequences to it.
I mean jesus christ, we are talking about something that affected childrenâs ability to formulate sentences correctly and their ability to read and write.
AI is at least a step up from that because you have to be able to articulate yourself properly if you want a proper conversation or answer.
So no. I am not really willing to talk about the âharms of using AIâ any more i am willing to talk about a campfire on the surface of the f-ing sun.
Just want to add that the dehuminizing concept doesnât work.
Chatbots are generated out of human text, and the generation is architected also by humans. Thus it can be only human.
But can a human or, say, group of humans direct how the chatbot decides when they generate it?
The answer is yes.
Otherwise, Nauseamedu is right.
The dangers in chatbots arenât being properly identified here, those problems are due to other things. The danger is rather the surrepticious coopting and redefining of what high intellect (is this not the position the chatbots hold in the popular mind?) is and what and how it thinks. Like they used to do with universities, but cheaper.
âI mean jesus f-ing christ, we are talking about something that affected childrenâs ability to formulate sentences correctly and their ability to read and writeâ
Yeah, but thatâs negligible. Werenât/arenât most middle-class millennials and gen zâers doing fine in school, getting jobs, buying shit, etc? If the effects of any social media and AI are deleterious, we should see it on a much larger scale. Look at the Japanese. Those people are the literal incarnation of video game people living in a digital media matrix. Walk around Tokyo, and all youâll see is creepy cartoon anime super heros on flashing billboards everywhere. They literally live in the Tiktok matrix. They are put on their first computers and iphones at age 2. A group of four Japanese adolescents will be walking through the subway and talking through a group chat app instead of just out loud to each other. Everything in the apartment is voice controlled.
My point is why are these people who are so immersed in tech not becoming zombies and imbeciles?
Ya know how itâll go down? Itâll take a thousand years, though. Robotics and AI will have literally made life so easy that critically important reasoning skills and such that you pick up from having to solve problems will be no more. Adults will be the mental equivalent of stoned teenagers. It wonât be like a Star Trek future where weâre all 150 IQs in perfect health and in light blue jump suits. Nope. The population will have been drastically reduced with machines taking over the workforce, and those people who do live will be like dopamine addicts on eternal vacation. A elite bourgeoise educated class bored out of its mind, so it starts using designer drugs. Think of Mingâs daughterâs ship with the sexy velvet padded chill room in the lounge. Thatâs what theyâll be doing. Orbiting erf in space ships in designer drug bars and kinky sex lounges. What else is there to do? Colonize Mars? Already done. Go back to erf where the robots and the zombie stoners are? No thanks.
Pretty sure. Its a question of how early you were introduced to it
The older you were before getting exposed to the miasma of bite size content, the less effect it has on you, though it still aint non negible.
As for the other part about life-style choices: Thats kinda a separate thing like any other âfanboyâ circle or whatever.
My point was more about actual physiological effects that trash has on people, especially children.
Its not like the topic of social darwinism has not been explored before.
And you cant exactly say that its wrong.
Instead of having natural selection, the human race is now busy with reversing it, saving every and any kind of genetically, mentally and/or otherwise flawed individuals, keeping them alive and letting them breed.
Sure.
You can make the argument that any part of your body which you do not exercise regularly, will eventually atrophy.
This includes the brain.
Is AI part of this issue?
Probably, unless it starts to take into account even this factor and gonna give you mental exercises.
But my point was mainly that we are already wrecking ourselves so hard that AI is basically not even qualifying as a nail in the coffin anymore.
Its just a shovelful of dirt thrown onto the casket we built.
