Victoria Harrington lives in South Paris.

AI, or artificial intelligence, is sold off to the public as something that is basically weightless; someone has thrown algorithms into “the cloud” that answers your questions instantly, something that can make money and something that can change the world forever.

This image doesn’t exist. AI runs on huge, energy-hungry data centers that are quickly changing local communities, the natural resources that are already straining, and it is also raising so many questions about the people benefiting from this “revolution” of technology.

As these places grow, so does the backlash. That backlash is completely justified and utterly overdue.

The amount of expansion is insane. To train these AI models, it takes immense computing power, way more than that of the internet. To meet the need, many companies are building these huge data centers that are filled with servers that flow 24/7. These machines can use as much electricity as a small city, and it can require just as much water for cooling purposes.

The places that trade off on these centers are increasingly questioning if it’s worth it. These centers are usually written off as economic opportunities, promising jobs, tax revenue and a great investment. These centers, though, are automated mostly so they don’t actually need a lot of workers to make them run smoothly, which means no long-term jobs after construction.

The people of the community are left to deal with the consequences. Higher electricity needs mean more strain on local grids, sometimes leading to higher bills and less reliability. The water usage is going to put a strain on the already dry conditions. In some cases, even, these centers could be taking away from housing facilities and small neighborhoods that could be used for more low-income housing.

The backlash that is happening isn’t about saying no to the technology. It is about the need for fairness. Communities are asking why, why should they have to bear the environmental and economic costs to something that basically only benefits corporations and distant users. When the water that is watering a rural area is rerouted for cooling systems for these machines, the imbalance is clear. When the people of the community are facing higher energy bills while these big companies get tax breaks, frustration turns to anger really fast.

There’s also a much bigger contradiction at play here. Many of the same companies driving the expansion of AI have made public commitments to sustainability and climate responsibility. But the fast-pace growth of these centers threatens to stop these goals.

Some of these centers are powered by alternative energy, but the overall need for electricity is rising dramatically and it risks outdoing the availability of the alternative power. In places where renewable energy is limited, it could possibly cause an increase to the reliance we have on fossil fuels.

This tension brings up another broad issue: the tech industry’s long-standing habit of letting others take the blame, pushing the downsides onto the people while reaping the benefits. For so many years, digital services have been known as low-impact compared to traditional industries. But AI? It is changing that.

The footprint on the environment that these centers are making isn’t good, and it’s becoming harder and harder to ignore. Communities are no longer willing to take this on without a question.

To address all of this backlash requires more than the public relations campaigns or tiny little improvements. It calls for a structural change.

First off, these companies need to be transparent by rule; it needs to be normal. Companies should have to disclose detailed information about their energy consumption, water usage and how its going to affect the environment at a local level. Without the data, the residents can’t make solid decisions or hold these corporations accountable.

Finally, the tech industry itself has to think about the difficult question: how much growth is enough? Everyone is in a race to build even larger AI models, a race that is mostly driven by competition rather than the need for it. Bigger is not always better, especially when the costs are being a burden on the residents who don’t even have a say.

AI may be our future, but it’s being built today — in actual places with actual consequences. The question is whether we are willing to see those consequences and do something about them, or whether we will continue to treat the cost of moving forward as someone else’s problem.