The literal cost of AI

Using GPT-4 to generate 100 words consumes up to 3 bottles of water — AI data centers also raise power and water bills for nearby residents.

According to a recent article posted yesterday by Christopher Harper, research conducted by the University of California, highlights the outrageous costs of using generative AI. ChatGPT requires substantial water consumption which is used to cool the servers that generate the data, and even small amounts, like generating text, can tax our resources, and in addition, severely affect the electric grid.

Data centers are shown to be heavy consumers of water and electricity, which also drives up the power and water bills of residents in the towns where these data centers are being built. For example, Meta needed to use 22 million liters of water to train its LLaMA-3 model — about how much water is needed to grow 4,439 pounds of rice, or, as researchers noted, “about what 164 Americans consume in a year.”

AI has been a great source of contention for numerous reasons, while many object its use with good cause, others tout it as the next best thing for mankind. But what are the benefits and risks to humanity, because let’s be realistic; nothing comes without a price.

The electric cost of GPT-4 is also quite high. If one out of 10 working Americans use GPT-4 once a week for a year (so, 52 queries total by 17 million people), the corresponding power demands of 121,517 MWh would be equal to the electricity consumed by every single household in Washington D.C. (an estimated 671,803 people) for twenty days. That’s nothing to scoff at, especially since it’s an unrealistically light use case for GPT-4’s target audience. 

The Washington Post included quotes from OpenAI, Meta, Google, and Microsoft representatives, most of which reaffirmed commitment to reducing environmental demand rather than giving actual courses of action. 

Microsoft rep Craig Cincotta stated that the company will be “working toward data center cooling methods that will eliminate water consumption completely,” which sounds nice but is vague on how. The AI profit motive has clearly taken priority over environmental goals set by these companies in the past, so even this quote should be taken with a grain of salt until we see actual improvements.