People on Lemmy love to cry AI IS BURNING THE PLANNET but have they ever really looked at the numbers?
- 1 kg Beef = 60kg CO2 - source
- 1000km Return flight = 314kg CO2 - source
- 1 Bitcoin transaction = 645kg of CO2 - source
- 1000 AI prompts = 3kg of CO2 - source including training
I’m a strong advocate for privacy, and I myself am Anti generative AI in many ways, but can we please stop shouting that AI IS BRUNING THE PLANNET, it weakens our position to push back at actual issues with AI like privacy.
Here is a detailed breakdown debunking that AI is destroying the planet: https://andymasley.substack.com/p/a-cheat-sheet-for-conversations-about
What I think this fails to account for is the most energy intensive part of the models is not interacting with them but training them, which uses exponentially more energy.
Edit: I read over your source more carefully and I don’t think the argument of “AI might be using 10 times less energy than we think” is an actual argument of any sort. I might have a golden pony who radiates pure love and shits chocolate donuts, prove I don’t.
Here’s facts, data centers are ramping up to train and operate AI and making water near their locations inaccessible source
Not to mention Musk’s gas turbines source
The carbon output from any form of large scale industry is going to have an impact but a larger issue for me is depletion of resources and a frenzied mania surrounding some imaginary ‘race’ to beat others at a technology which is yet to turn a profit for anyone but Nvidia
By your numbers 1 prompt generates 3g of CO2
ChatGPT says they handle 2.5billion prompts per day.
That’s 7,500 tonnes of CO2 per day from ChatGPT alone. The vast majority of those are completely pointless queries.
Your point is spurious at best and misleading at worst.
I would be curious to compare that to Google search. Some experts estimate Google search and the time spent reviewing the results on your device can be 3-10g of Co2. The Google part alone is maybe .2g per search.
https://www.fastcompany.com/90171268/internet_impact_visualized
Here they estimated in 2018 that just rendering front page cost 300 tons a minute, so 432,000 tons a day. Just for the front page.
Just to put it in perspective that the Internet itself is not free. It’s expensive to power sending all this data around!
Yes, when you go by billions, the numbers quickly add up.
For beef, you can feed 1 KG to 3 people, which creates 60 KG of CO2.
If you calculate this for Europe, we’d get about 14900000 tonnes of CO2 PER DAY.
So promting, even from the biggest LLM currently on the planet, is still only 1/2000 of eating beef in Europe alone for 1/3 of a single meal.
Your point is misrepresenting even more, and using statistics to blow up the number, while not doing the same calculation for the other parts.
So ChatGPT isn’t as polluting as the beef industry, what’s your point? We can go on asking for LLMs to make pictures of big tiddy anime girls and the answers to simple questions that it still gets wrong because people eat meat still?
Just like bitcoin and all the other pump and dump cryptos, LLM chatbots created a market that didn’t exist before that needlessly adds excess pollution (and to be clear, I agree with OPs primary point that the environmental impact isn’t the worst thing about them), and they’re trying to force them into every aspect of our lives whether they’re suited to it or not.
Highlighting the environmental (and commercial, social and privacy) concerns is one tool to get policy makers to maybe think twice about signing another huge deal to get ChatGPT embedded in yet another part of our lives.
It’s highlighting hypocrisy. It’s asking: do you take this problem seriously, or are you just complaining?
Having LLMs shoved into everything is a serious problem. But it’s a problem the way that forced updates and invasion of privacy were already a problem. Fixating on energy use is pretense. It’s working backwards to point at the negative externalities of something you’ve already made conclusions about, as if those factors were relevant to your conclusion. Using that as rhetoric is the nature of bad faith.
My point?
If you complain the world is burning because of AI but you eat beef, you are contributing >2000 times as much as a non-beef eater who uses AI. Making you a hypocrite who is mad at the wrong things.
Neither AI not beef are a necessity.
This is exactly my point
You’re completely missing my point.
Streaming 4k video uses a lot more, should we also critique people who stream video?
1 Bitcoin transaction = 645kg of CO2
I think it’s interesting that this very much depends on the coin. This article compares Bitcoin to Monero, and here are the figures (per transaction per year):
- Bitcoin - ~95 TWh per year
- Monero - ~1.5 TWh per year
So Monero is ~63x more efficient than Bitcoin at transactions, and that’s with all the overhead Monero bakes in to its transactions to maintain privacy (each transaction generates a bunch of fake transactions to mislead snoops, and it’s notoriously ASIC/GPU inefficient).
I’m guessing the same is true for different AI models, some models will be a lot more wasteful than others in their training and queries.
I agree, the opposition to LLMs is misplaced. There are legitimate reasons to dislike it, and there are certainly policy changes we should make, but attacking it for energy use in the weakest way to oppose it. If LLMs ran on 100% green energy, would you still oppose them? Probably, so figure out why you oppose it and attack LLMs over that, and perhaps policy can alleviate the worst of those concerns.
Thanks, this is exactly my point
The shipping industry emits a billion tons of CO2e per year. Training a model emits… maybe a thousand? An impact that could be offset by reducing Chinese imports by 0.0001%. Or arbitrarily limited by strong-arming the very few companies involved. DeepSeek knocked off a few orders of magnitude and R1 seems to work, as well as any of these things work.
But some people don’t really give a shit about the electricity involved - it’s just a negative for them to latch onto.
Now, it is a problem locally, where datacenters turn power straight into heat, as fast as they can manage. Anything like a tiny sliver of the global shipping industry becomes noticeable when it’s concentrated in one building within range of a commute.
Hmm, after running an LLM locally I’m not too surprised about low energy use while querying (but even then, 3Wh is very low, I was getting like 5 Wh per query while running DeepSeek locally, that said datacenter GPUs will be more efficient for this). But also, this doesn’t go into great depths about the impact of training.
It doesn’t go into the impact of training, because that isn’t what these AI-phobes complain about.
Even just using it will result in calling it “slop” and complaining about its energy usage.
If they complained about the training, they would have a point. But they don’t. They complain about anything of AI. Even if it has revolutionized things like protein folding in medicine.
My sources account for the training.
Even if you add a zero to AI’s CO2 emissions and water usage, it’s still a drop in the bucket compared to eating beef or taking a flight for holiday
Check the sources, that does account for training.
Even if we went with 10wh, it’s still a drop in the bucket compared to other lifestyle choices we make
Check the sources, that does account for training.
The 3Wh figure is pretty much “pulled out of the ass”. Looking at other sources listed, it does seem true that similar-sized models use ~2-3Wh per query, amortized training included. So yeah I concede the point.
Upvoted for controversial post…
This whole analysis is based on the cost of queries, which consume, in fact, a small amount of power.
You (or OOP) also acknowledge that the massive, obscene, megalomaniac power consumption of AI comes from building up the models themselves.
And then you proceed to ignore that the first one is the driver for the second. This is asinine. It’s exactly like saying that the CO2 emissions of fossil fuels is the cost of running the pumps at the station.
Here you go, according to the same logic:
Look at my sources, it accounts for the training of the AI etc.
And all the power spent by tons of companies now constantly scraping the internet?
And all the storage now that all the big players need to have local copies of the internet to train on?
All the power spent making all those hard drives? All the power it takes for Nvidia to make millions more GPUs than otherwise needed?
And the power their employees need to drive to work?
And the power for all the computers they used growing up?
And the power to make those computers?
And the power Intel used inventing the microprocessor?!
And the power the entire telephone grid used while AT&T developed the transistor?!?!
Still a drop in the bucket compared to people taking flights to go on vacation.
If you’re not angry at people taking flights, you shouldn’t be angry at AI for environmental reasons.
There are plenty of other reasons to hate AI.
I agree that crypto and casual air travel are also wasteful and that there are plenty of other reasons to hate AI.
But it’s reasonable for someone to believe that recreational travel is beneficial in a way that AI isn’t. I could do without either, but there’s nothing hypocritical about thinking that two very different things have different cost/benefit calculus.
deleted by creator
