🤷 How can baseless claims about AI spread so widely?

Over the past two weeks, I have shown that reporting claiming AI data centers consume enormous amounts of water and significantly raise temperatures is incorrect. The claim that has gained the most traction is that AI is highly energy-intensive, which is also false.

Despite not being true, these claims have spread widely in both traditional and social media and had a major impact. There isn’t a single Q&A session during a talk where someone doesn’t ask me about this.

How is it that something so wrong can gain such reach? Is it a conspiracy?

A narrative is more powerful than a conspiracy

No, I believe it is something else.

Claims about AI did not spread because they were well-founded, but because they landed right in the middle of two strong currents of opinion at the same time.

One was the climate debate, where the need to reduce dependence on fossil fuels is genuine and important. The other was the concern among journalists and creators that AI will replace them. That group is particularly aware of and concerned about climate change.

Here came something — AI — that threatened both the climate and their own jobs.

And, as we know, the media is always especially quick to highlight negative news and warn about dangers.

The claims about AI took on a life of their own. When enough people repeat something, it starts to feel true, even to those who should know better. Those who pass it on have rarely done the groundwork. They believe it because so many others have said it, which makes even more people believe it for the same reason. It is not a conspiracy, but when this happens, it is more powerful than conspiracies.

I fell for it

I fell into the trap myself. For quite some time, I believed that AI was especially energy-intensive. When asked, my response was that it was worth it—that we could use AI to reduce energy consumption in other areas.

However, some people began to scrutinize these claims. One of them was Hannah Ritchie, a climate activist who once believed the world was heading in a terrible direction until she saw Hans Rosling’s TED Talk and realized that progress and suffering coexist. She changed her path, became a researcher at Oxford, and began working at Our World in Data, a site that presents statistics on global development.

In other words, she is someone who cares about the environment and climate, wants to see a transition away from fossil fuels, and has the knowledge to think independently and resist strong narratives.

She wrote both that the claims about AI and energy were not true and placed energy use in context. One strength of the claims about AI’s energy or water use is that most of us have no real sense of how much energy is used in other parts of our lives.

When I read her texts and examined the sources, it began to dawn on me that I had probably been misled. While working on my book about AI, I dug into the numbers myself and realized that I—and almost everyone else—had been wrong. AI was not particularly energy-intensive.

But even after realizing this, it felt wrong. Despite the facts, I felt like a “climate denier” when answering questions during talks—as if I were trying to downplay a drawback of a new technology. It is bizarre to feel that way, but that is how it was. That feeling has now passed and instead turned into anger over how, once again, baseless claims are holding us back.

The tide is starting to turn

No matter how strong narratives are once they take hold, they are not immovable. Hannah Ritchie herself began studying the claims after reading Andy Masley’s review. He is an independent researcher and writer who is now beginning to gain global attention.

More and more people are picking up his analyses and helping to spread them. Previously, there was not even a response—now there is, and I believe I can see that the tide is beginning to turn.

Despite this, several major news outlets should be embarrassed. It is unacceptable that you are spreading this kind of baseless propaganda.

Mathias Sundin
Angry Optimist