DEV Community

Cover image for The Obedient AI God That Devours Our Resources
OLEH POLISHCHUK
OLEH POLISHCHUK

Posted on

The Obedient AI God That Devours Our Resources

OLEH POLISHCHUK
Independent Senior AI Logic Auditor & Forensic Investigator | AI Digital Transformation Researcher & Publicist | Author of the Concept of Reasonable Sufficiency™ & Developer of the Forensic Logic Auditing (FLA) Method™

Silicon Valley marketers sell AI as an "ethereal, cloud-based, nearly holy salvation for the planet." We are promised that algorithms will optimize energy grids and save the climate. But this is the greatest civilizational self-deception: humanity thought it had created a manageable deity to solve all problems, but in reality, built a blind, gluttonous idol.

Thesis 1: The Myth of "Pure Reason" AI does not exist

as a whole natural or artificial phenomenon. It is a marketing simulacrum. What they call artificial intelligence is a fragmented, distributed network of a colossal number of servers, cables, and other people's mathematical code. It has no agency or "mind" that cares about ecology. It is simply a giant piece of physical infrastructure that desperately wants "to eat" and right now "devours" our very own non-infinite and even scarce natural resources.
Billions of people have been offered to unconditionally accept a name that has no relation to intelligence in the human understanding. People have no representation of what this actually is, therefore, based on the name, have imagined possibilities and capabilities for this imperfect phenomenon that it does not possess. In this mass self-deception, they convinced themselves or accepted on faith the necessity of enduring all critical negative consequences of this phenomenon, reconciling with them, closing their eyes to these consequences, and demanding this from others who do not agree. For what?

Thesis 2: International Sabotage Under the Guise of Progress Humanity

has ready-made mechanisms at its disposal, but they are not observed. The Paris Agreement is openly ignored. The USA and China — the two largest economies in the world — are the largest sources of natural pollution. And this is despite the fact that both countries annually suffer from the heaviest destructive natural cataclysms. But even this — billions in damages to infrastructure, private property, and finally, human deaths — does not force the governments of these countries to start saving their countries, to change their attitude towards ecology.

Tell me, does the presence of multimodal AI models in the phones of billions of people somehow have the capacity to increase the defense capability of any country, affect crop yields, change the direction of monsoons, etc.? I think not. Just a few years ago, before the appearance of AI, we were quite normally making online orders, payments, hanging out on social networks, and much more. At the same time, millions of people had jobs. Why is this not being talked about today as something that we have lost? In exchange for what?

And now data centers are being added and added to this equation. Can anyone clearly answer how many such centers need to be built, where, how many resources will they consume and, most importantly, for what? How does all this harmonize with the Paris Agreement? Not at all. The digital world has hit physical limits of microchips, rare-earth metals, wires, transformers, and water.

One modern AI data center with a capacity of 100 MW consumes:
Electricity: about 876,000 MWh per year — like 100,000 European homes.
Water: up to 3.6 million liters of clean fresh resource per day simply evaporates into the sky to cool the silicon. And this is in conditions of the growing deficit of this resource, which is crucial for the survival of civilization.

Billions of dollars that should go to the modernization of national energy grids, the creation of energy storages, and the real development of renewable sources are disappearing into the "digital furnace." Silicon Valley investors are simply taking the ready infrastructure, which was built for decades for people, and are redirecting this resource to service their commercial AI models. The most cynical part is that during power shortages, they are transferred to "backup sources." And what are backup sources for giant server rooms? These are, as a rule, most powerful, dirty diesel generators or old gas/coal thermal power plants (TPPs). As a result, under the sauce of a "green digital future," companies at moments of peak loads start burning fossil fuels on an industrial scale, poisoning the very atmosphere that is already heated to the limit.

And how exactly do these centers in the end "help" ecology? To "forecast" global warming, they accelerate this very warming many times over, right here and now. This is pure logical paradox and managerial madness!

Thesis 3: The Deadly Compromise We have run into a systemic crisis.

A situation when in the most technologically developed economy in the world one has to choose whom to supply with power in forty-degree heat — the air conditioner in an American's home so that they do not get heatstroke (or in hospitals), or another farm with H100 chips that generates tons of images and compiles texts — this is exactly the systemic crisis and dead end. When AI makes a logical mistake, it does not just set back a business — it burns planetary resources in vain.

I can write a book that will consist only of answers from leading AI models, where they hallucinate, make mistakes, admit this themselves, and endlessly apologize for their imperfection, which leads to exhausting but necessary checks of their answers, to the inability to perform the required work. And this, I want to note, is not related to searching for answers to how many Martians died during the colonization of Jupiter. I want to repeat once again what I said in all my previous articles — I am not an opponent of progress. I am for progress, but not at such a price!

Conclusion for B2B and Compliance:

Ignoring this is no longer possible. In the coming years, an ESG audit (ecological compliance) will start to severely strangle corporations for the use of "dirty" and gluttonous models. If a business right now does not start conducting an independent logical audit — which tasks really require AI, and where it is simply a waste of resources for the sake of fashion — companies will inevitably get burned on environmental fines, lawsuits, and reputation.

Top comments (0)