The Gist
- AI's efficiency is misleading. While AI can produce content rapidly, the environmental impact of this "efficiency" is devastating.
- Environmental costs of AI. The energy, water and materials required for AI operations far outweigh any perceived efficiency gains.
- Long-term consequences. AI’s resource-intensive processes are contributing to significant environmental degradation, challenging the true meaning of efficiency.
AI is so much more efficient than a human at producing content, we’re told. AI writing a page of text emits hundreds of times less CO2 than if you got a human to write that page. AI creating an image can emit thousands of times less CO2 than a human creating an image. This is how the business case for AI goes. Fire the writers, fire the graphic artists, fire the programmers.
The Efficiency Illusion: AI vs. Human Content Creation
We know the old saying: “Sorry, I wrote you a long letter because I didn’t have time to write you a shorter one.” Quality writing, great artistry, it’s nearly never about quantity. It takes time to do things well.
Sure, AI is fast and furious. Sure, it can churn out a page of text or an image much faster than a human can. It can churn out tens of thousands of pages and images in the same time as a human creates one. And it does.
Flooding the internet with advertising and marketing spam, political misinformation, fake novels and biographies, fake, error-riddled scientific papers, horribly written computer code. And on a page-by-page basis, on an image-by-image basis, it’s efficient. It’s super-efficient at filling the internet with crap in super-quick time.
Related Article: From Trash to Misinformation: AI's Journey
The Hidden Environmental Costs of AI’s Rapid Production
Then … Then … In the blink of an eye, AI is gobbling up an India of electricity, gurgling down a South Korea of water, crapping mountains of mining waste and valleys of e-waste.
Gobbling, gurgling, crapping, lying … And … And … Wait for it. When they’re challenged about this incredible environment-destroying waste, the tech bros’ response is: Yeah, these chips and servers and data sets and coding and algorithms are not efficient “yet.” We’re innovating towards efficiency. Next model, for sure, we’ll be more efficient.
Hold on! You said a moment ago you were so much more efficient than humans, and now you are saying that this super-scary doubling of electricity, water and materials, this warpath you’re on to destroy the environment, that it reflects inefficiency, and that if we just keep going on this dystopian AI warpath, the AI advertising war machine will become more efficient?
Something doesn’t compute, AI. Either you’re efficient or you’re not efficient. This technological efficiency sure is killing us because efficiency has never meant what we’re sold that it means.
Related Article: How AI Bias Creates Dependency and Inequality
The Environmental Price of AI's ‘Efficiency’: A Dystopian Reality
In one hour, a skilled artist can create one image. Leaving aside quality for a moment, in that hour, AI will have created one million images. That’s what they mean by efficiency. Calculating the cost of an individual image unit created by a human versus an image unit created by AI, AI wins on cost and efficiency, no questions asked.
However, the total cost to the environment for AI for its million images churned out is vastly greater than the cost of the human effort. And the total long-term costs of mining waste, energy pollution and e-waste of AI are vastly greater, vastly greater.
In this finite world of limited resources, of a super-stressed environment, AI and Big Tech “efficiency” is literally efficiently killing us, efficiently killing our environment. It’s an efficiency scam of monumental, dystopian proportions.
Learn how you can join our contributor community.