Anonymous #1 7w

"AI is going to save us and needs to be willingly adopted" and "AI is going to kill us and needs to be banned" as blanket positions are the main ones I see, even though there is so much nuance in between those two extremes given that AI is not one fixed thing. Some forms of AI are already saving us whereas other forms are already killing us. It's not a matter of AI vs no AI. Studies have shown AI outcomes are a matter of which form and how it's used.

10

Anonymous #2 7w

Your reasoning leans on a lot of generalizations, so let me break it down. First, AI isn't one thing. It's a class of technology with wildly different forms, uses, and ethical implications. Not all AI lives in data centers, and data centers don't only run AI. Data centers ARE the cloud: Google Drive, Spotify, Netflix, YouTube, Instagram, basically anything you log into. Amazon alone hosts a huge chunk of the internet, and self-hosted apps are rare these days.

6

Anonymous #2 7w

Second, AI is only one slice of that load. Video streaming and social media are massive consumers, and a typical chatbot session is small next to hours of doomscrolling and binge watching. On water, data centers are dwarfed by agriculture (beef especially) and lawn irrigation.

5

Anonymous #2 7w

Third, the part that actually deserves anger is how cloud AI is financed. Chip and cloud companies invest in AI labs, and the labs spend that money buying those same companies' chips and cloud capacity. The money loops through a small group of firms, which makes demand look bigger than it is and inflates valuations. Hence the bubble. A lot of the harm people blame on "AI" is driven by that capital race, the same profit engine behind most ecological destruction, not by the math itself.

2

Anonymous #2 7w

That doesn't mean there's no harm. Dropping frictionless chatbots on an unprepared public had real costs. While research suggests heavy cognitive offloading can dull thinking, it also shows experts who engage critically with LLMs produce better work. It cuts both ways.

5

Anonymous #2 7w

That's why I think people need critical thinking and prompt literacy first, and why local, self-hosted models are a better path: they take real technical knowledge to run, which is a natural barrier to entry, and they pull power away from the cloud giants. Same with owning your media like local files, CDs, DVDs.

3

Anonymous #2 7w

But "get rid of AI"? How? It's math. Open-source models are already downloadable and run on ordinary hardware. A ban would be nearly impossible to enforce, and history (alcohol prohibition, cannabis) shows bans push things underground and make them less safe, while regulation actually reduces harm.

6

Anonymous #2 7w

You can't ban a movie, but you can rate it. You can't erase an equation, but you can regulate the industry, crack down on the financing games, and give people real alternatives to channel their dissent into.

5

Anonymous OP 6w

nobody just says “ai will kill us”. they say “if we keep racing forward, recklessly, and build ai smarter than we are that we don’t understand, and hand over more and more power to it, it might eventually kill us”

5

Anonymous #3 6w

and these people are often experts in the field!

6

Anonymous #3 6w

I was not talking about experts, I was talking about the layman. The layman anti-AI position is not nearly as nuanced, since the average person is informed by mainstream media, which is optimized for engagement and thus systemically prone to misinformation and fearmongering. You're right that officials say racing forward recklessly in AI development could kill us, but think about what that implies and who benefits from us believing that...

5

Anonymous #3 6w

Racing forward is not a quality of AI, it's a quality of capitalism. I personally believe massive AI compute in datacenters is a huge mistake given that AI's complexity provides a natural computational delay, a period of time where you have to sit and wait. Slowing down and being methodical is exactly what we are missing under late-stage capitalism, and the shortening of that delay is driven by the status quo, not by the math.

1

Anonymous #3 6w

Now think about what happens when you internalize the belief that AI is going to kill us if we keep developing it. The average person doesn't understand that AI can run locally, so they see "AI" and datacenters as being the same thing. This makes AI into a metaphysical thing, which puts distance between its governance and its users, just like the internet at large.

6

Anonymous #3 6w

Most people doing know anything about computer science, and it makes implementing a solution feel urgent because the user is made to believe they lack control, and urgency is precisely what got us into this mess in the first place. Because most people don't understand how the cloud works, or the internet for that matter, they don't realize how tied it is to capitalism and so they urgently yet blindly attack the wrong things.

0

Anonymous #3 6w

This is what bothers me about the average AI discourse. It's typically generalized, lacking intellectual rigor and accuracy. People believe what the officials say without questioning their motives. A narrative of destruction and power benefits the elites immensely because it gives them justification to act urgently. This is why we typically hear that narrative from people like Sam Altman. It's no coincidence.

0

Anonymous #3 6w

They know power is not something that can be created, it must be coerced, and the way it's stolen is through systemic obfuscation and hyper convenience. This dynamic of systemic disempowerment is by no means exclusive to cloud AI, and that's why it bothers so much when people attribute it to AI. It's a much bigger mechanism that has been around for a very long time. AI is merely illuminating this pre-existing dynamic.

1

Anonymous #3 6w

So in essence, AI is not solely responsible for the harm we see and speculate. Profit motive is.

0

Anonymous OP 6w

*don't

0

Anonymous OP 6w

*bothers me

0

Anonymous OP 5w

Omg you’re so psychologist rn you’re being so neutral and using profound language it’s great I’m proud of you thank you for going to school , fellow psych student

6

Anonymous OP 5w

i mean yeah powerful and misaligned ai is the direct risk/harm and the indirect thing that could lead us there is profit motives and coordination problems. but i don’t see how saying this helps the ai companies. “if you allow us to just act from capitalism’s incentives without regulating us, we might all die” how is this supposed to help them get more profit?

2

Anonymous #3 5w

Here's how it helps them: if AI gets framed as one godlike thing that only a few labs can handle safely, the "solution" becomes licensing and compliance costs that only billion-dollar companies can meet. That's a moat. "Our product is so powerful it could end the world" is also great for valuations. Meanwhile the AI regular people could actually own, small models running on your own computer, gets lumped in with the datacenter stuff.

0

Anonymous #4 5w

Thank you! That really means a lot bc I love psych but I'm actually not primarily a psych student 👀

0

Anonymous #3 5w

"Alignment" isn't neutral. Somebody decides what counts as aligned, and right now that's a handful of companies with financial incentives. I'm not against AI being safe. I'm against a few corporations getting to define "safe" for everyone and calling it science. The question isn't whether to align AI, it's who gets a say.

0

Anonymous #3 5w

I'm not saying nobody sincerely worries about misalignment. Lots of researchers do, including ones who don't profit from it. My point is narrower: "aligned" always means aligned to someone's goals, and right now the ones deciding are the same companies racing each other. That's what I want people questioning.

1

Anonymous OP 5w

i mostly agree but those 3 replies are 100% ai generated as detected by me and then pangram 😭

4

Anonymous #3 5w

You want a cookie? I stand behind everything I said because this is MY position, and attacking the tools I used to present my own genuine thoughts doesn't in any way refute them. Try again.

-1

Anonymous OP 5w

ah shit, I was agreeing as well until this point If you made a poster or something, that’s using tools to present your thoughts. Asking AI to write your responses is just finding a close enough approximation of your thoughts. Even if it’s correct, you’re just painting your own argument in a worse light for no reason, especially in cases where AI use is literally the topic of discussion.

1

Anonymous #5 5w

in hindsight I don’t know how I didn’t see it earlier, you literally said “so let me break it down”.

1

Anonymous #5 5w

Refute my points, not my methods.

0

Anonymous #5 5w

you agreed with it when you read it. the words didn't change. the only thing that changed is you found out how I wrote it. if the argument was good enough to convince you an hour ago, tell me what's wrong with it now. otherwise you're not disagreeing with me, you're just embarrassed you agreed.

0

Anonymous OP 5w

I’m not refuting your stance, I agree that there’s more nuance than “ban AI”. I’m not even disagreeing with you. I’m stating my own stance, which is “AI has places it is useful and places where it is harmful. Using it as a replacement for your own thoughts makes you look dumb and/or dishonest”.

0

Anonymous #5 5w

And I disagree with that framing. Research shows cognitive offloading in novices resulted in mental atrophy, while it also shows experts who engage critical produced increased quality of work. It cuts both ways. Using LLMs does not at all automatically imply “replacing your own thoughts.” That’s disingenuous of you to insinuate.

0

Anonymous OP 5w

*critically

1

Anonymous OP 5w

I literally just said I don’t disagree with the argument? I don’t understand why you’re assuming I do. you, or whatever you used, mentioned the same points I would have made. I am saying that I am disappointed by your methods, find them ironic to the topic at hand. Our disagreement is only whether using AI to “offload cognitive effort” is a good or bad usecase.

1

Anonymous #5 5w

You’re right that we’re not disagreeing about that. But you can’t criticize my thoughts by how they are conveyed. That goes for any conversation, AI or no AI. Cognitive offloading is a feature of every information processing system. That’s like the entire point. It’s a substrate. The problem is not simply offloading cognition, it’s when cognition is unconsciously offloaded. If you want to talk about unconscious cognitive offloading, short-form social media is a much bigger culprit.

1

Anonymous #5 5w

It’s our responsibility as thinking beings to make sure the information we are taking in is accurate. That’s the deeper issue.

1

Anonymous OP 5w

I’m not really understanding what you’re trying to say here? social media isn’t a tool to accomplish a goal, it just feels like a reward for an accomplishment to your brain. That’s a completely different issue from people letting AI accomplish something for them that they’re meant to do themselves. I wouldn’t call using social media cognitive offloading, unconscious or otherwise.

0

Anonymous OP 5w

Your generalization about information systems depends how you define “cognitive offloading”, but I would disagree. Writing a response is completely within what you can do as a person, and doing it with AI is a shortcut. The information processing your phone is doing every second exceeds what can be done by a person. Using it isn’t just offloading doing all the networking and graphical maths by hand, because that wouldn’t be possible in the first place. it’s creating an entirely new experience.

0

Anonymous #5 5w

Yes, I use AI to help write these, and I'm going to keep doing it. I'm neurodivergent. My thoughts loop and branch, and when I write unfiltered it comes out sharper and more condescending than I mean. AI helps me turn ideas I've worked through for years into something people can actually read. That's an accommodation, not a replacement for thinking.

0

Anonymous #5 5w

On cognitive offloading: it just means handing mental work to something outside your head, like notes, calculators, calendars, or search engines. The research doesn't say offloading is bad. It says outcomes depend on whether you stay actively engaged. That's my whole point: using AI doesn't automatically mean not thinking. And you've assumed what I believe a few times now. If you want to disagree with my position, ask me what it is.

0

Anonymous #5 5w

I don't mean scrolling social media is itself offloading. I mean what happens when people take in a reel’s claims without checking them. The algorithm decides what you see, the creator decides how it's framed, and you absorb the conclusion without doing the evaluating yourself. That's handing off your judgment, and most people don't notice they're doing it. My point is that unconscious outsourcing like that is a much bigger, older problem than someone knowingly using AI to help them write.

1

Anonymous #5 5w

And I'll be honest: I'm genuinely fucking offended. You called me dumb and/or dishonest for using a tool that helps me communicate from a brain that works differently than yours. You didn't ask why I use it; you assumed the worst and judged me for it. That's not a neutral take on AI. It's a judgment about how I'm allowed to think and speak, and that’s a dick move.

1

Anonymous OP 5w

just type your replies normally i don’t get the point

3

Anonymous #3 5w

I just explained the point above. You don't have to get it, but I'm not going to stop using a tool that helps me just because you'd prefer I didn't.

1