Use of LLMs in Academia
Hey all, I just want to offer my perspective here as someone who uses AI for work, and why the anti-AI crowd makes me extremely nervous.
I do physics research at a state university in the US. Last year, our physics department head sent out an email to everyone encouraging us to use AI, talk to AI, use it for our research. Why? Because physicists like tools that make our jobs easier. We have left wing and right wing professors here and they all use AI.
I was scared to use LLMs until around 2025 when I got that directive. I didn't touch LLMs until recently, and I have to say, LLMs are smarter than the average physicist if you use them correctly. A mediocre physicist + LLMs can make the contributions of a much cleverer scientist. Autonomous laboratories and autonomous mathematicians are real. They are already here. So I can firmly say that the misconception that AI is 'dumb' is flat out wrong.
I do find it hilarious in a terrifying sort of way that the anti-AI crowd uses the fascist playbook here. They claim that AI is both weak and strong. Simultaneously a threat and completely worthless. Don't buy it. AI is real. Academia knows it. The government knows it. Corporations know it. AI is here to stay. Shaming each other for using it isn't going to make the big bad AI un-invent itself.
Pandora's box cannot be unopened. Shaming people for accepting that is a completely self-destructive enterprise. All you are doing is creating artificial division that should be reserved for hating capitalism. If we could just redirect everyone's anger from AI to capitalism, we just might achieve class consciousness. The closest we got was with the epstein files, but I think that has finally fallen out of the news cycle. We won't get that energy back by hating AI.
What do I use it for? I use AI to teach myself skills that broaden my resume. I have strengthened my technical skills and broadened my horizons. I use AI to write analysis code faster. I use it to help understand scientific papers outside of my field. I use it as a study partner. I use it to quickly cross-reference documents. My group mates use Gemini to build analysis tools.
The anti-AI crowd seems to think that the only possible way to engage with an LLM is to just turn your brain off and let it think for you, instantly frying your brain and killing you. This terrifies me because it smells of the 'But how do atheists have morals if they aren't afraid of eternal hellfire?' Instead it's 'But how do the AI users stop themselves from cheating if they don't hate AI?' All you're telling me is that if YOU used AI, YOU WOULD USE IT TO CHEAT. Just like the christians are actually telling me that if YOU didn't believe in eternal damnation, YOU WOULD MURDER PEOPLE. Your failure to take pride in your own work isn't my problem. I use AI. I take pride in my work. I don't let AI do my job for me. I use it to teach myself and get things done faster. The same is true of any productivity tool. Like a cell phone, a pencil, or a car.
We were told that if we don't use AI, we are handicapping ourselves, because a scientist + LLMs can basically double their workload by using AI. Previously, a tool that would have taken two weeks to code can be generated in an afternoon. Are we cheating? Of course not, no more than it would be cheating to use a computer instead of handwriting your thesis with pencil and paper.
But I've seen comedians that I like shitting on AI during crowd work. Someone will say they use AI for work, and the entire crowd collectively sighs. If the anti-AI crowd knew that STEM research is AI-assisted, they would lose their shit.
I don't even pay for chatGPT. I use the free version. I'm not even giving Sam Altman money by using his product. When I talk to chatGPT for free, I cost Sam Altman money. I am bankrupting him, and you can too. So I know that argument doesn't work. Sure, I'm complicit in the violence caused by data centers poisoning their environment. But as long as everyone has given up on regulation, they are also complicit. I say regulate it, obviously. The harms can be mitigated, just like with any other technology. And that leads me to my most speculative point.
I am terrified that the Anti-AI panic is a capitalist psyop.
Just think about all the people who hate AI. All the people who demonize others, shame others. All the vitriol and anger in their hearts. Aimed at each other. It makes me sick. All of that rage should be directed toward capitalism. But instead, everyone has chosen to divide themselves even more. We were so close to class consciousness. I'm disgusted that it slipped through our fingers, and now the so-called 'left' is engaging in such witch hunts.
I think there is one extremely important reason why everyone is so angry about this: Because it's the only power they think they have. We can't put trump behind bars for the epstein files. We can't stop fascism. We can't stop capitalism. We can't get zionist money out of politics. We can't get health care. But you know what we CAN do? Shame each other for using AI. We can start a witch hunt. We can attack each other. Because that's possible. But stopping capitalism? Fixing the real problem? Not possible. So people naturally slide down the slippery slope into witch hunts. I'm not letting any of them off the hook. Shaming each other is taking the easy way out and capitulating to the perceived inevitability of capitalism.
Big shout out to Jlyplaylists in the comments for linking this video, but basically yes, billionaires are funding this propaganda. The point of it is to make AI feel bigger and more terrifying than it really is so that everyone is more scared of AI than of capitalism. Truly a brilliantly cynical plan. They are encouraging the very hatred that puts themselves in the sights of all the rowdy people. But the catch is that they know they're perfectly safe.
They're perfectly safe because everybody is angry at AI instead of capitalism. And they will remain perfectly safe until we tear down capitalism. AI is not the enemy. Capitalism is.