All views shared in the Opinions section of the Chatterbox belong to their respective authors, and may not represent the views of the publication as a whole.
Happy Saturday! I hope you all are having a restful weekend and are getting a lot of homework done. These days, it can be extremely tempting to use artificial intelligence (AI) to check homework answers or to make a study guide. Before you do so, I would like to share some information about the consequences, specifically for the environment.
I feel like we’ve all been told by teachers that AI is “bad for the environment,” but sometimes the information on how exactly is lacking. For this reason, I decided to do a deep dive into the impacts of AI.
It’s easy to think that when you put a simple search request into AI it has no impact. However, just like many things these days, one action can create a long-lasting, multiplying butterfly effect of ecological harm.
For one, AI has been found to increase greenhouse gas emissions, something we can not afford since our Earth is nearing the precipice. To quantify this, here are some comparisons borrowed from Massachusetts Institute of Technology (MIT):
One search in a generative AI tool requires four to five times the energy needed from a regular search browser, like Google.
One image generated by AI uses the same amount of energy as charging a phone to max battery.
Training one AI model consumes five times the energy an average American car uses throughout its lifetime.
These numbers illustrate the huge amount of energy that AI requires, most of which comes from fossil fuels, producing greenhouse gases. AI also uses an immense amount of water. The computers used to run AI get extremely hot, more so than a typical data center, and require freshwater to cool them down.
One large AI data center can use 5 million gallons of water per day, which is how much water a town of 10 to 50,000 people uses daily.
Even when you do a search using AI, you are consuming water. An estimate from the University of California, Riverside found that a 100-word search query uses around one bottle of water. Multiply this by the billions of AI users, and you will begin to see the problem.
Most of the water used to cool data centers is freshwater that comes from local aquifers and water sources. As a result, data centers put a huge strain on community water supply, which is especially consequential in areas such as the West that are already experiencing drought.
All of this energy and water consumption isn’t just bad for the environment; it’s also bad for your wallet. The expense from energy and water used by data centers gets placed on consumers. Electric bills are already up 6-14%, depending on where you live; this will only increase as more people use AI.
The computers and processors used to run AI also have some detrimental environmental consequences. If you read my article on electric vehicles, you will remember me explaining how the materials used to create rechargeable batteries are often mined unethically and in a wasteful manner.
The same is true for AI hardware; many rely on rare elements and minerals that are not mined in a clean way.
This is not to say that AI does not have its benefits. AI does have some good attributes, so I thought that I would highlight them so that you can ensure you have a well-rounded opinion.
AI can be used to track emissions, model weather patterns and provide quick disaster response. AI was also used recently by MIT to find a greener alternative for concrete, which is a very carbon-intensive building material. Finally, companies such as The Ocean Cleanup have used AI to accelerate the removal of plastic from oceans.
All of these examples illustrate how AI can be used in a positive way. However, it is important to realize the extent to which we should use AI.
In my opinion, AI is valuable for compiling data and doing things that would be far too time-intensive for a human to do.
However, it needs to be used sparingly. Don’t use AI to complete tasks that you can do in five minutes. AI shouldn’t be used to generate pictures of unicorns on motorcycles or posters for your club meeting.
You can also remove the automatic AI tools that come on search engines, for example, the Google AI Summary. This is a great way to cut down on AI reliance for menial tasks. Learn how to delete it here.
Use AI as a tool, not a crutch; try to rely on it as little as possible. Opt for non-generative forms of AI, since these use less power than their generative counterparts.
There are also actions that large companies need to take. AI tech companies need to implement more green infrastructure and methods in data centers, perhaps by recycling water, reusing computer parts or investing in renewable energy.
All of this to say, the next time you go to use AI for a simple search request, think about the consequences this action will have. I hope you will see that AI has a time and a place and that we need to be more considerate with how we choose to use it.
Elena