Can I ethically use LLMs?
Can I ethically use LLMs?
The title is not a rhetorical question, and I'm not going to bury an answer. I don't have an answer. This post is my exploration of the question, and why I think it is a question.
If you mean from an energy/climate/water/resource consumption perspective, then no. But of you're looking at it from a labor perspective, then also no. From a copyrights perspective? Nope as well. Okay, but surely from a correctness perspective? Very clear no. Okay, but there's still the aspect of showing recipients respect and not wasting their time by giving them something to read/view/process that you didn't care to write/think through yourself in the first place? Well, you guessed it, hard no as well.
The things that AI was made for are:
- finding a use for superfluous Blockchain GPUs when that hype began to die (initially), and more importantly...
- replacing human workers who demand nasty things like wages and vacation and sick days and rights, in order to redistribute wealth to the wealthiest.
copyright is immoral
Yes, because private property is theft. But unequal enforcement of copyright law is worse. Right now, LLMs are just lying machines trained on pirated data and the companies that run them are acting with impunity for doing something a normal person would get put in jail for.
Copyright is immoral, but as long as it exists, the laws should be extra strict on companies that steal others' works.
I'm fairly certain blockchain GPUs have very different requirements than those used for ML, especially not LLMs. In particular they don't need anywhere as much VRAM and generally don't require floating point math, nor do they need features like tensor cores. Those "blockchain GPUs" likely didn't turn into ML GPUs.
ML has been around for a long time. People have been using GPUs in ML since AlexNet in 2012, not just after blockchain hype started to die down.
In the same way as people can be okay or not with eating meat, I think it comes down to individual values.
LLMs have consequences to climate, labour, and (if you don’t disclose in some situations) how others perceive your creativity.
For me, I would have no problem asking and AI agent to speak to a colleagues AI agent to automatically find a mutually acceptable time for a call or meeting.
I would not use AI to compose a best man’s toast or funeral eulogy from scratch.
The quote about “I want AI to do my dishes and laundry so I have time for more creativity, I don’t want AI to do my creativity so I have more time to do dishes and laundry” applies to me.
LLMs have consequences to climate
I would have no problem asking and AI agent to speak to a colleagues AI agent to automatically find a mutually acceptable time for a call or meeting.
Yeah I mean the climate's important but screw talking to people
I eat meat, too.
Like, vegan "nothing with a face!" ethically? No. Absolutely not.
Vegetarian "I eat eggs and dairy?" Sure.
I think you can if you do not involve it in creating something.
For example, giving it a document and asking it questions about it.
The only "creation" I've tried to do is a cover letter for job hunting. If I needed something other than a resume, portfolio AND references, then I see nothing wrong with using one to do work that another ai is going to look at... Just my $0.02
By "creating something," do you mean replacing the work that a human creative would normally do? (That is, replacing an artist/writer/musician who would otherwise have the job?)
That's where I'd draw the line. I think personal use is fine but if you're affecting people's employment by using an AI that's clearly a no-go for me.
I'm unsure that using complicated algebra to regurgitate parts of people's works is different from just copying it. Perhaps you could say a human brain learning how to code is just regurgitating the code it's had as an input before, but intuition says directly copying is somehow different.
I add copyleft licenses to code to ensure the code cannot be legally copied into proprietary software, for moral beliefs. If the output of a LLM was free software and copyleft (as would be the input) then perhaps that would be fine. Github probably has some complicated legalese that says by uploading it you permit them to use it for LLMs - I'd want that to be legally voided.