AI and Sustainability: Navigating Reality

 (Source: Designhubz.com)

Last week, I promised to use this week’s blog to show the results of an initial exploration of the role that AI presently plays in sustainability. Sonya Landau raised the question with me, so I used Copilot (Microsoft AI) to look into three important, sustainability-related themes: global human survival and sustainability, energy cost, and sustainable reality. As I said last week:

Meanwhile, the Sustainability program/course is excellent preparation for the students’ post-undergraduate lives. By the time present students graduate, it is probable that they will be able to generate content for a sustainability course through AI. The challenges will be to motivate the students to go beyond this and to give them the tools to critique the generated content.

Below are the results of my three searches:

“Global human survival and sustainability”

Copilot

Sent by Copilot:

Ah, the grand challenge of our time: ensuring the survival of both our species and the beautiful blue marble we call home. 🌍 Let’s dive into this, shall we?

1. Planetary Boundaries: A Wake-Up CallScientists have been busy with their cosmic stethoscopes, and they’ve diagnosed Earth with a few health issues. You see, our planet has these nine primary “planetary boundaries” that act as vital signs. Think of them as Earth’s version of cholesterol levels or blood pressure. And guess what? We’ve been a bit reckless.

• Climate Change: Yep, we’ve blown past this one. Our carbon emissions are like a fever that won’t break.
• Biosphere Integrity: That’s a fancy way of saying biodiversity. And guess what? We’re not doing so hot. The Amazon rainforest side-eyeing us right now.
• Freshwater Availability: Imagine a water cooler at a global office. Well, it’s running dry. We’re guzzling more than we should.
• Land Use: We’ve been redecorating the planet, but not in a good way. Urban sprawl, deforestation, and highways—oh my!
• Nutrient Pollution: Picture phosphorus and nitrogen as the party crashers in Earth’s aquatic ecosystems. They’re causing algal blooms and ocean dead zones. Not cool, phosphorus. Not cool.
• Novel Entities: These are like the uninvited guests at the planetary party—microplastics, radioactive waste, and other human-made pollutants. They’re littering the place.

But wait, there’s hope! Ocean acidification, air pollution, and ozone depletion are still within the safe zone. So, it’s not all doom and gloom. These boundaries aren’t “tipping points” of no return; they’re more like flashing warning signs. Like when your blood pressure spikes—you know you need to cut back on the fries and take a walk. Earth’s giving us side-eye too, saying, “Fix this, humans!”

2. Sustainable Population: How Many Is Too Many?Now, let’s talk numbers. How many of us can comfortably share this cosmic apartment? Well, estimates vary, but recent science suggests we might be overbooked. Like a crowded elevator during rush hour.

• Some say the sweet spot is between 2 and 4 billion people. Yep, that’s quite a range. It’s like saying, “Party’s full; no more RSVPs!”
• Paul R. Ehrlich, a population guru, thinks the optimum population is even lower—around 1.5 to 2 billion. That’s like having a cozy dinner party with close friends.

So, what’s the prescription? Well, we need to be mindful roommates. Sustainable living, conservation, and responsible resource use are our best bets. Imagine Earth as a co-op—everyone pitches in to keep the place tidy. No leaving dirty dishes in the sink (looking at you, plastic waste).

3. Human Sustainability: It’s PersonalBut here’s the twist: Our survival isn’t just about Earth’s health. It’s also about our well-being. When we’re stressed, burned out, and running on fumes, we’re not great roommates. So, self-care matters too. Think of it as maintaining your own little corner of the planet.

Remember, we’re all in this together. Earth’s our shared home, and we’ve got to be better tenants. So, let’s turn off the lights when we leave a room (literally and metaphorically), recycle, hug a tree (figuratively, unless you’re a koala), and maybe—just maybe—consider carpooling to the cosmic block party. 🌱🌎

And hey, if you ever need a pep talk about planetary boundaries or sustainable living, I’m here. Just like a cosmic life coach, minus the crystals and incense. 

The main references, based on which this AI gave this response, are shown below:

Smithsonian: Humans Have Exceeded Six of the Nine Boundaries Keeping Earth Habitable

Wikipedia: Sustainable population

Thrive Global: Introducing Our Human Sustainability Hub

Springer Link: Creation, humanity, science and sustainability for human survival

My own descriptions have been a bit more modest. Table 1 in the August 20^th blog shows the global changes in five key areas that took place over my lifetime in the 10 most populated countries. These changes directly impact our global environment, with some of them having a direct impact on global survivability. This theme repeats in number of blogs.

“Energy cost of AI”:

Ah, the energy appetite of artificial intelligence—quite the hot topic these days! 🌍🔌

So, let’s dive into it. AI, especially the more powerful and resource-hungry models, can indeed be quite the energy hog. Imagine a model generating an image—just one image—using something like the Stable Diffusion XL. That seemingly innocent image creation process can gobble up as much energy as charging your trusty smartphone. 📱💡

But wait, there’s more! If you were to churn out a thousand of those images, you’d be emitting as much carbon dioxide as driving just over four miles in a gas-powered car. 🚗💨 Now, I don’t know about you, but that’s a lot of virtual road trips for some pixels!

However, not all AI tasks are equally voracious when it comes to electricity. Creating images? Energy-intensive. Generating text? Not so much. It’s like comparing a marathon runner to someone casually strolling through a park. 🏃‍♂️🌳

And here’s the twist: using smaller, task-specific AI models can be dozens of times more efficient than summoning the colossal, all-purpose ones. It’s like opting for a compact car instead of a gas-guzzling SUV. 🚗🌿

Now, let’s zoom out a bit. The big picture: electricity consumption from data centers, AI, and cryptocurrency combined could double by 2026, according to projections. Yep, that’s right—AI, data centers, and crypto are collectively adding “at least one Sweden or at most one Germany” to global electricity demand. 🌐🔌

So, while AI is flexing its computational muscles, it’s also doing a little dance with our power grids. But hey, it’s not all doom and gloom. Researchers are working on making AI more energy-efficient, like teaching it to sip electrons instead of chugging them. 🤖🍵

Remember, though, these numbers aren’t just about AI; they’re a mix of data centers, crypto mining, and AI. It’s like trying to figure out which ingredient makes your grandma’s secret sauce so addictive—tricky business! 🤔🔍

In the end, balance is key. We want AI to be smart, but not at the expense of our planet. So, let’s keep innovating, optimizing, and maybe throw in a sprinkle of solar panels for good measure. ☀️🌎

And hey, if you ever need a virtual energy-saving tip, just holler—I’m here! 😊👋

Again, the main three references, based on which this AI gave its response, are shown below:

MIT Technology Review: AI is an energy hog. This is what it means for climate change.

Nature: Generative AI’s environmental costs are soaring — and mostly secret

Penn Today: The hidden costs of AI: Impending energy and resource strain

Here is another recent opinion on this: MSN: The AI boom is spurring an energy crisis: is there a sustainable way out?

Essentially, there are a lot of cases where AI is being used unnecessarily. Separating the tasks that need AI from those that don’t will cut back on energy use. Also, once electricity generation is fully carbon-free and mostly solar (see the June 25, 2024 blog), AI’s energy hunger will be considerably less threatening.

“Sustainable Reality”:

“Sustainable Reality” often refers to the concept of integrating sustainability into various aspects of life, including investing, real estate, and business practices. One notable example is the Morgan Stanley Sustainable Reality report, which highlights how sustainable funds have outperformed traditional funds, especially during volatile periods like the COVID-19 pandemic¹². This demonstrates that incorporating environmental, social, and governance (ESG) criteria can potentially provide better financial returns and reduce investment risks. 

This AI description reads more as advertisement for Morgan Stanley than a productive summary of the issue.

Ideally, I would have liked to change the title of the required, college-wide category of the ASU General Education course that was described last week’s blog. I would have re-named it “sustainable reality,” similar to the title of Roger Penrose’s book, “The Road to Reality,” with the obvious difference that Penrose focused on Einsteinian Cosmology, while my focus is confined to our human reality.

Admittedly, all three AI descriptions that I am including in this blog are initial responses and represent me opening a conversation with the AI site. I have asked students in the past to start such conversations in my cosmology course, starting with large abstract questions (see “How to Explain Reality,” January 26, 2022).

I will return to the role of AI in future blogs to follow the progress of the technology. Even the best AI will not be very helpful with navigating our present reality. The simple reason is that the all-important training of our computers that run the AI is based on present and past realities that are in a constant state of change.