The operation of data centers to process artificial intelligence is projected to use a huge amount of electricity and water for cooling. How many gigawatts of electricity is iNaturalist using for its operation? I hate to think that iNat is causing nuclear power plants to be built. Or a solar power farm in what is a pristine desert ecosystem.
5 Likes
You may be interested in this thread and those linked in it:
https://forum.inaturalist.org/t/carbon-footprint-of-inaturalist/11126
10 Likes
I dislike AI for the most part, but if AI is what drives a switch to nuclear power, then Iām all for it. Nuclear is likely the only realistic form of clean energy that can support the demands of the power grid
45 Likes
From the post that @thomaseverest linked to:
10 Likes
A few points here:
- I could be wrong, but Iām pretty sure that iNatās CV AI is a lot less intensive to run than todayās state-of-the-art giant LLMs.
- iNatās user base is large, but still orders of magnitude smaller than major AI platformsā (~300k active users compared to ~400m for OpenAI, ~20m for Claude, ~30m for Microsoft Copilot, etc.)
- This wasnāt really your main point, but nuclear power is very much a net good for the environment (compared to other sources of power, including solar)
32 Likes
From an environmental perspective nuclear power is absolutely the way to go. Vastly less impact than pretty much every other power source. The anti-nuclear crowd and the fear around it has done immense harm in forcing societies to stick with polluting, damaging, and dangerous ways of generating power, and has massively stifled safety and efficiency innovation in nuclear power.
38 Likes
Not all AI is equivalent to generative AI, and not all AI (or things that are now being called AI due to the current hype around it) is bad. Obviously, there is some processing power and server power required for the CV to work, and to host all the images, etc. And quantifying exactly how much that is would be pretty difficult, and is not within my scope of expertise.
But to say iNat is ākilling the planetā is almost certainly not true. Itās of course good to be critical of that in a world where it seems everything is being pushed towards automation despite the ecological costs, but the good that iNat does should be factored in as well. Especially when compared to generative AI that steals art, makes up information, and plagiarizes real people while also taking up a ridiculous amount of space and energy.
29 Likes
Personally I am in support of nuclear power, but I still donāt like the idea of nuclear power plants being built solely for AI. And that is happening, for example a nuclear plant near my hometown is reopening solely to provide power just for a single companyās AI. Iād rather we didnāt keep adding to the energy demand and that the nuclear power plant could instead replace other more harmful power plants for a net positive. Even renewable power sources have environmental footprints, so in my mind itās just adding to the existing harm.
And, unfortunately, AI is also driving the construction of gas-powered plants and other non-renewable energy production. There are plans to build the nationās largest gas-powered plant in my state to meet the energy demand of AI. Under current plans, the plant would be one of the biggest sources of carbon pollution in the state once it opens.
But I really donāt think that iNaturalistās impact is in any way comparable to the impact of language models like ChatGPT.
12 Likes
I think the environmental impact of AI is somewhat overblown, too. There are many other things that are far, far more of a problem.
8 Likes
This question is kind of assuming that people would be doing something less consumptive if they werenāt using inat, and I donāt think I would.
12 Likes
While I agree that there are bigger problems that need to be addressed, I would suggest that, if anything, the environmental impact of widespread AI use has been massively underreported and that itās a far larger issue than the majority of even decently well informed people realize.
Just looking at power use for processing, and nothing else, the AI industry uses more electricity than almost all other countries does for the entire nation. And thatās not including what goes into the infrastructure development in the first place.
Globally, the electricity consumption of data centers rose to 460 terawatts in 2022. This would have made data centers the 11th largest electricity consumer in the world, between the nations of Saudi Arabia (371 terawatts) and France (463 terawatts), according to the Organization for Economic Co-operation and Development.
By 2026, the electricity consumption of data centers is expected to approach 1,050 terawatts (which would bump data centers up to fifth place on the global list, between Japan and Russia).
Once a generative AI model is trained, the energy demands donāt disappear.
Each time a model is used, perhaps by an individual asking ChatGPT to summarize an email, the computing hardware that performs those operations consumes energy. Researchers have estimated that a ChatGPT query consumes about five times more electricity than a simple web search.
Chilled water is used to cool a data center by absorbing heat from computing equipment. It has been estimated that, for each kilowatt hour of energy a data center consumes, it would need two liters of water for cooling, says Bashir.
- https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117
- https://mit-genai.pubpub.org/pub/8ulgrckc/release/2
And thatās not even getting into the various other issues surrounding it.
20 Likes
Although Iām not sure how realistic this is, I am hoping that generative AI will just be another āfadā that people (and the tech industry) will eventually move on from. Pretty much everyone I have heard from on the subject is tired of it being constantly shoved down our throats by every tech company under the sun, and frustrated at the amount of artificial slop on the internet that has resulted from its adoption (if you have ever looked at youtube trending videos you know what Iām talking about, not to mention all the fake art, music etc). As a software dev it can be mildly helpful to me when looking up usage examples of more obscure coding concepts, but beyond that I have never had any use for it besides making chat GPT write ridiculous stories/generate ridiculous images for kicks and giggles (which I got bored of very quickly). And I think most people feel a similar way
8 Likes
I suppose this could happen, but I doubt it. I think it will gradually become integrated into our everyday lives (like the internet has become) and we will simply get used to it, but I donāt think people will get bored of it.
2 Likes
I think virtually everything we humans do damages the environment in some way. It becomes a question of how much damage vs. how much gain and vs. whatever weād do if we werenāt using iNaturalistās CV. A very important point is that iNaturalistās CV uses much less energy than most of the AI forms we hear about because it examines a much smaller base of data and it produces a much more limited output (just a choice from a set of names).
There are benefits to the environment from our using it. Do the benefits (including helping people interact with nature) exceed the costs? Hard to quantify. Iāll say, I hope so. Maybe they do. Iāll keep using it.
9 Likes
Like other replies said, real problem is generative ai. iNat CV have a clear purpose, itās for protect nature and biodiversity. but generative ai is far from thatā¦ itās so bad. https://en.wikipedia.org/wiki/Artificial_intelligence#Ethics
Why people only see good sides (convenient) but ignoring all of bad sides? I donāt get it.
6 Likes
Normally I just read here, but as I am reading a lot of ānuclear for the environmentā, I want to add sth to think about:
While nuclear energy is often praised for its low-carbon output and potential to meet large-scale energy demands, itās important not to romanticize it as the ultimate environmental solution. The reality is far more complex, and any serious discussion must consider the full pictureāenvironmental, economic, technical, and societal.
From a climate perspective, nuclear power may perform well. It has a low carbon footprint over its lifecycle, comparable to wind and lower than solar, and it provides a stable, dispatchable energy source. In theory, this makes it valuable for supporting baseload energy needs, especially in a future with growing electricity demands driven by AI, digital infrastructure, and electrification.
However, this does not automatically make nuclear the ābestā or most realistic solution. Nuclear energy comes with significant drawbacks: long construction times (often over a decade), extremely high upfront and total costs, and unresolved issues around long-term radioactive waste storage. Moreover, the environmental consequences of uranium mining, thermal water pollution, and the catastrophic riskāhowever rareāof accidents cannot be ignored.
While anti-nuclear sentiment has played a role in slowing its development, itās inaccurate to blame public fear alone. Energy systems have remained fossil fuel-based due to a complex mix of infrastructure inertia, political and economic interests, and in some cases, poor policy planning. At the same time, solar and wind technologies have advanced rapidly and are now the cheapest, fastest-growing sources of electricity in the world. Of course those come with their own challengesāintermittency and land useābut these are increasingly solvable with modern storage solutions, grid improvements, and smart energy planning.
In short, nuclear can play a role in a clean energy future, particularly in countries with existing infrastructure or limited renewable options. But itās neither a silver bullet nor a universally applicable solution!! The science and economics show that a sustainable energy transition will likely rely on a mix: primarily solar and wind, supported by storage, demand flexibility, andāwhere feasible and affordableāsome nuclear.
The debate around nuclear energy isnāt about being for or against it. Itās about recognizing that energy systems are complex, and climate solutions need to be fast, scalable, economically sound, and socially acceptable. Simplistic narratives, whether pro- or anti-nuclear, donāt help us get there.
:)
13 Likes
Nuclear? And radioactive waste in perpetuity?
https://e360.yale.edu/features/us-uranium-mining-nuclear-power
https://ejatlas.org/print/nuclear-waste-storage-near-the-spanish-frontier-of-portugal
Cape Town has Koeberg. In the city near residential housing. No ābunkersā. Evacuate on busy roads?? Built by the French, similar plants in France are shut down as they have reached the end of their designated life span. Ours is having its life extended, and the 2 units take it in turn to go down for planned or UNplanned maintenance. Previous manager emigrated to Canada. Staff turns over.
We have solar, and wind, and molten salt batteries for storage. What we lack is transmission lines.
I would favour microgrids to generate and supply the power where it is needed.
https://mybroadband.co.za/news/energy/590897-eskom-becoming-an-internet-provider.html
4 Likes
Well, ācleanā is when you donāt look at the tons of radioactive waste.
How about a Dyson sphere? Sure, extremely difficult, futuristic, but canāt we?
1 Like
Youāre absolutely right that itās an intractable question to answer in any way that takes everything into consideration. I spend a lot of time on iNat, and for sure that comes with an environmental impact; but if iNat didnāt exist, I would obviously spend that time doing something else. Who knows if Iād have another hobby like painting or motorcycle racing (with its own environmental cost) if all my iNatting time were to be freed up? Looked at this way, iNat is only damaging to the environment if its average environmental impact (per user-hour) is higher than for the average human activity. These kinds of numbers are impossible to calculate with much accuracy.
Nope.
iNaturalist does not reduce genetic diversity.
1 Like