I was having lunch with a friend, talking about professional tennis, when we wondered about the relative heights of tennis legends Nadal, Federer, and Djokovic. I reflexively picked up my phone and spoke the question to ChatGPT. In seconds we had the answer.
When my trail cameras recently captured images of birds that I didn’t recognize, I showed them to ChatGPT for identification. It told me what they were.
And just before writing this, I wanted to invite a friend to a lunch celebrating his birthday. Instinctively, I picked up my iPad and asked ChatGPT to create an image of a birthday cake with lettering in the icing that said, “Happy Birthday Rick.”
This is my AI life now. Whenever I have a question, an answer is close at hand. Whenever I want an image, ChatGPT or Copilot (free) is eager to create something delightful.
What’s not to like? Just ask my colleague David Fisher, who has been nagging me to write about the environmental issues related to artificial intelligence. Okay, maybe that’s not a nice way to say it. Let’s just say that he’s been sending me emails emphatically encouraging me to write about this.
Another email just came: “Don’t want to spam you but: ‘According to a report by Goldman Sachs, a ChatGPT query needs nearly 10 times as much electricity as a Google search query.’”
And there’s the email he sent me a week ago with the subject: “AI is unconscionable.” He wrote, “As a matter of conscience, I won’t knowingly use ChatGPT or any other AI-dependent technology. . . . You could at least give this all-too-serious concern a good strong sentence or two in one of your columns. Any claimed benefits of AI are far outweighed by the environmental costs—and will become ever more so in the foreseeable future.”
He followed up with an email quoting The Guardian: “According to the International Energy Agency, total electricity consumption from data centers could double from 2022 levels to 1,000 TWh (terawatt-hours) in 2026, equivalent to the energy demand of Japan, while research firm SemiAnalysis calculates that AI will result in data centers using 4.5% of global energy generation by 2030. Water usage is significant too, with one study estimating that AI could account for up to 6.6bn cubic meters of water use by 2027—nearly two-thirds of England’s annual consumption.”
Water? I was curious why AI requires water, so naturally I asked ChatGPT. The response was: “[W]ater is used in the cooling systems of data centers where AI computations are performed.”
Then I wanted to know why the water couldn’t simply be recycled, and ChatGPT explained that too. The water used accumulates impurities, and it’s complex and costly to treat this water. Plus, treating and recycling water requires energy and infrastructure, adding to the cost. Along with four additional reasons.
What’s the solution to these environmental costs? Pull the plug on AI? I guess we know that that’s not going to happen. And even if the plug were pulled, the energy consumption of huge data centers would still be seriously problematic. A typical data center, according to Perplexity AI, is hundreds of thousands of square feet containing many thousands of servers. Google alone has several dozen data centers.
Who uses these data centers? You do. Almost everything we do now involves data centers. David sends his emails via Gmail, which is hosted in Google’s data centers. As I write this column, each time I save the changes, that information is uploaded to a data center. When I shop on Amazon, I’m accessing information stored in their huge data centers.
An article on this topic in The Atlantic was titled, “Every Time You Post to Instagram, You’re Turning on a Light Bulb Forever.” The author talks about the energy it has taken for a data center to store a 15-second video he posted on Instagram nine years ago. He writes, “Every time we make a new video or send an email, or post a photo of our latest meal, it’s like turning on a small light bulb that’ll never be turned off.”
It’s a situation we need to reckon with.
I guess my feeling is that this is nothing new. We’ve been barreling toward an uncertain future ever since the evolution of human intelligence, language, agriculture, cities, civilization, writing, printing, the industrial revolution, electronic media, computers, and now AI. Historian Yuval Noah Harari says we are the ultimate invasive species, infringing on every ecological niche.
But what I see is a long history of emergent intelligence, the spontaneous expression of an intelligence that underlies all of existence. And I feel that ultimately, AI will force us to come to a deeper understanding of the nature this intelligence—and consciousness.
I love AI not only for its utility, but for what it may ultimately reveal about ourselves.
Find column archives at JimKarpen.com.