The impacts of artificial intelligence on environmental sustainability and human well-being
This systematic review of 1,291 studies reveals that while AI research often portrays environmental impacts as predominantly positive, it suffers from narrow methodological scopes and an uneven assessment of human well-being, which shows mixed outcomes across different dimensions, thereby highlighting an urgent need for more comprehensive, empirical, and holistic evaluations to guide AI toward true sustainability and human flourishing.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine Artificial Intelligence (AI) as a massive, lightning-fast new engine that has just been installed in the car of human civilization. It promises to take us to amazing new destinations faster than ever before. But, just like any new engine, we need to ask two big questions: Is it polluting the air we breathe? (Environmental impact) and Is it making the passengers happy and safe? (Human well-being).
This paper is like a team of mechanics and sociologists who looked at over 1,291 reports (studies) to see what everyone is saying about this new engine. Here is what they found, translated into everyday language:
1. The Two Different Stories
The researchers noticed that the people studying the engine are telling two very different stories.
The Environmental Story (The Mechanics): Most of these reports are very optimistic. They are like mechanics who only look at how much fuel the engine saves. They say, "AI helps us use less energy and cut carbon emissions!"
- The Blind Spot: They are mostly looking at the engine's immediate fuel consumption (energy and CO2). They are ignoring the mining for the metal, the water used to cool the engine, the trash from old parts, and the fact that if the engine gets cheaper, we might drive more, not less (a "rebound effect").
- The Analogy: It's like praising a hybrid car for saving gas, while ignoring the fact that the battery factory is draining a local river and the car is so cheap that everyone buys a second one, clogging the roads.
The Well-being Story (The Sociologists): The people studying how this affects people are much more mixed and worried. They are split almost 50/50.
- The Good News: AI might make us richer and help doctors diagnose diseases better.
- The Bad News: AI might make inequality worse, cause job losses, spread lies (misinformation), and make us feel lonelier or more anxious.
- The Analogy: It's like a new smartphone app that helps you find your way (good for health/income) but also steals your data, makes you feel lonely because you talk to bots instead of friends, and takes away your job as a taxi driver.
2. What We Are Missing (The "Invisible" Parts)
The paper points out that we are looking at the engine through a very narrow keyhole.
- Ignoring the "Upstream" Mess: We talk a lot about what the AI does (application), but almost nothing about where it comes from.
- Analogy: We love the smooth ride of the car, but we ignore the fact that the rubber in the tires was harvested by underpaid workers in the Global South, and the mining for the copper wires destroyed a local forest. The paper says we need to look at the entire supply chain, from the mine to the trash heap.
- Ignoring the "Systemic" Changes: We focus on small improvements, but we aren't looking at how the whole system changes.
- Analogy: If AI makes traffic lights smarter, we save time. But if that saved time makes people decide to drive 50 miles instead of 10, we end up with more pollution. This is called the Rebound Effect. The paper says we need to study these big, long-term shifts, not just the small wins.
3. The "Feeling" Gap
The researchers found a strange split in how people feel about AI:
- Environment: 83% of studies say "Great! It's good for the planet!" (Mostly because they are looking at efficiency).
- People: It's a toss-up. 46% say "It's bad for society," and 44% say "It's good."
- The Catch: Even the "good" feelings are specific. We think AI will help our wallets and health, but we are terrified it will hurt our jobs, our trust in each other, and our sense of fairness.
4. The "Black Box" of Our Brains
There is a new, scary worry that hasn't been studied enough: What is AI doing to our brains?
- Analogy: If you use a calculator for every math problem, you eventually forget how to do math. The paper suggests that if we use AI to write, think, and remember everything for us, we might start to lose our own ability to think deeply, remember things, or feel emotions. We are "offloading" our brains to the machine.
5. The Big Picture: We Need a New Map
The authors propose a new way to look at AI. Instead of looking at the environment and society separately, we need to look at them together.
- The Old Way: "Does this AI save energy?" (Yes!) -> "Great, let's build it."
- The New Way: "Does this AI save energy, but at the cost of water, mining rights, worker exploitation, and increased loneliness?" -> "Maybe we need to slow down and fix the rules."
The Takeaway
AI is a powerful tool, like fire. It can cook our food and keep us warm, but if we don't control it, it can burn down the house.
The paper concludes that we are currently too focused on the small wins (efficiency, speed, profit) and ignoring the big costs (inequality, environmental damage, mental health). To make sure AI actually helps humanity flourish, we need to stop looking at just one part of the picture. We need to study the whole journey—from the mine where the metal was dug, to the factory where it was built, to the way it changes our jobs, and finally, how it changes the way we think and feel.
In short: We need to stop asking "How fast can AI go?" and start asking "Where is it taking us, and who is getting left behind?"
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.