# My Turn | Artificial intelligence and credit cards: Who pays the debt? - The News-Gazette

*Источник: The News-Gazette*
*Дата: 2026-09-29*
*Язык: en*

**Кратко:** I use artificial intelligence — or AI, as it’s commonly known — and I also use credit cards. A chatbot tells me that AI and credit cards are both “something people find incredibly powerful and thrilling initially, only for the true costs to reveal themselves much later.”
Half of credit-card holders carry a balance from month-to-month, and in the U.S., the average credit-card user is about $7,700 in debt, or a total of $1.6 trillion, according to the Federal Reserve Bank of New York.

I use artificial intelligence — or AI, as it’s commonly known — and I also use credit cards.
A chatbot tells me that AI and credit cards are both “something people find incredibly powerful and thrilling initially, only for the true costs to reveal themselves much later.”
Half of credit-card holders carry a balance from month-to-month, and in the U.S., the average credit-card user is about $7,700 in debt, or a total of $1.6 trillion, according to the Federal Reserve Bank of New York.
I’m not saying eliminate credit cards, or AI, but use them responsibly.
The benefits of AI are clear — at least to me, a professor and a scientist. AI has already changed the way I do work: with AI, I just assume everything can be found, so I concentrate on what isn’t known.
Designing new specific and small antibodies in the computer and then testing them out experimentally is just one example in Alzheimer’s research, a field I work in.
When I recently planned a visit to Europe, I asked AI to make an itinerary for me. It did it for free, costing me about $2,500 in hotels. A human travel agent I consulted was asking $7,000.
But where is the cost of AI?
The difference between how I use my credit card and AI is that I pay off my credit card every month, so the bill doesn’t grow.
Two of the hidden costs of AI are in the electricity and water needed to run the vast data centers.
Consider electricity: In Ohio, for example, 100 miles outside of Cincinnati, Nvidia, OpenAI and SB Energy are planning the world’s largest data center. It would use an amazing 8 to 10 gigawatts of electricity — more than enough to run every single household in all of Ohio and equivalent to eight to 10 full-scale nuclear reactors.
Instead, to get things up and running by 2028, they are building a massive natural-gas power station on the site. Elon Musk, the world’s richest man, and his SpaceXAI company are building a similar site in Tennessee and Mississippi — again, by building on private lands. These states are not alone. Virginia, a “leader” in data centers, uses 40 percent of its electricity for data centers.
These companies are in a race to get to be the first “intelligent superhighway” for AI. Every month of delays, according to Morgan Stanley, costs them $1 billion in missed revenue.
But in their rush, they are burning fossil fuels to get there. Some 56 percent of electricity used to power data centers nationwide comes from fossil fuels, according to the Environmental and Energy Institute.
And you are paying for it. Ameren (and ComEd) have recently raised rates by about 30 percent, and they say that’s because their costs have gone up. They certainly have! Ameren buys electricity in bulk that comes from the whole Midwest. So, Illinois is, in effect, paying for the data centers built in Ohio.
This is a reality even though Illinois has passed a no-new-data-centers initiative, which appears to be spreading across the country.
And with burning more fossil fuels, the environmental consequences are real — even if not completely immediate. Earth’s average temperature has already increased by 2.4 degrees Fahrenheit, largely due to humans’ burning of fossil fuels.
And it’s getting worse: the recent catastrophe in the Himalayas is a dramatic example of what will happen next. In the Northern Hemisphere, the Arctic is losing permafrost, and as it thaws, microbes have more to eat, creating yet more carbon-dioxide emissions, which will cause more ground to thaw.
Water is also a concern, as the centers use vast amounts to keep their chips cool. Yet about two-thirds of planned data centers are to be built in drought or water-starved areas. Why? Cheaper land and tax incentives.
Instead, cooling could be done by using enclosed water systems and the vast heat-sink of Earth itself. Why not do this? It would make AI more expensive. Instead, it’s cheapest to take our water and just run the wells dry.
Although this outcome could become our reality, we can slow it down, both through our local utilities, our universities, our high-technology companies — and exercising some political will.
Utilities such as Ameren were given a monopoly on all things electric in exchange for treating the electricity like the utility that it is and giving investors a reasonable rate of return. Meanwhile, Ameren’s CEO got a pay raise of 45 percent last year — earning $14.1 million plus $200,000 in airfare. This expense is justified by the company’s “strong financial and operational performance,” which includes a 50 percent increase in the stock.
It’s like having another utility — e.g. public schools — raise money by issuing stocks, with the school principal paid with stocks instead of being based on the quality of students and their graduation rates.
And where are the universities?
“We are working on clean-energy solutions toward the goal of not impacting the environment,” says Dean Rashid Bashir at the Grainger College of Engineering here at the University of Illinois. This includes new micro-nuclear reactors, a new seminar series entitled “Energy-Water-Compute,” and working with the National Science Foundation.
But there is an incredible impetus to hire more AI gurus rather than AI people whose focus is ameliorating the cost of AI. Meanwhile, the NSF, while once thought of as pursuing “unbiased research,” now must have AI front and center or it isn’t even considered for funding.
Nationally, why can’t the government make a law that says that AI (or any) companies can’t use any more fossil fuels or water? Musk says the entire U.S. electrical network could be run by a 100-mile-square array of solar panels. Experts agree with this (to a factor of two). And could you imagine what motivation for solar power there would be?
Meanwhile, our own president has paid over $4 billion to cancel offshore wind turbines that were on the verge of making electricity in an environmentally friendly manner.
And does it really matter to you if Google, OpenAI, Anthropic or any other company gets to open the AI floodgates first? Why can’t the government force them to cooperate instead of competing, slowing down the AI juggernaut, at least until they have solved the problems of electricity and water?
Yet, like a kid who is offered candy — or an adult with a credit card — the appeal of AI is too great, with payment delayed.
Someday, I will happily use my AI, knowing that I’m not building up a massive “debt” that my kids and grandkids will be paying off.

[Оригинал](https://www.news-gazette.com/opinion/guest-commentary/my-turn-artificial-intelligence-and-credit-cards-who-pays-the-debt/article_02ccdc98-b1ef-48ce-9543-d0acd9e9ccd1.html)