AI: Appetite For Injury

Please help support Pollutical.com!

In the most recent Pollutical post, Banks Expect A 3° World, we saw that one of the reasons decarbonization isn’t happening—apart from carbon offsets being largely a gigantic fraud—is that newly commercialized AI, artificial intelligence, requires large, power-hungry data centers.

There’s nothing inherently wrong with using AI—it’s hugely beneficial, even transformative, in medicine, for instance. The Temerty Faculty of Medicine at the University of Toronto literally has it in their slogan: Transforming health through AI.

1. “Revolutionizing healthcare: the role of artificial intelligence in clinical practice,” Alowais, S.A. et al. BMC Med Educ 23, 689 (2023). DOI: 10.1186/s12909-023-04698-z

2. “Overview of artificial intelligence in medicine,” Amisha et al. J Family Med Prim Care. 2019 Jul;8(7):2328–2331. DOI: 10.4103/jfmpc.jfmpc_440_19.

But in a profit-driven capitalist system, its primary uses aren’t determined by a cost-benefit analysis, like new diagnostic techniques vs. new emissions, or growth-conscious strategies that look at what current emissions could be substantially cut back to make room for large-scale AI analysis of anonymized public health data to identify hidden patterns. Instead, to a large extent, it’s driven by businesses trying to apply AI to get an edge on competitors and pay less to employees, military contractors looking for lucrative new avenues for profit-seeking, and students using it to cheat on term papers while surfing social media.

In many cases, it seems, companies that were happy to go along with bullshit carbon offsets if it would polish their trashy image aren’t about to cut back on investments in AI even though that, too, would be a good look. There’s simply too much money to be made. Which means much larger energy bills, payable in the form of increased climate emergency disasters.

Statista.com recently released a series of graphs that show where we’ve been and where we’re going.

For one example, let’s look at Meta’s capital expenditures, recently ballooning due to AI. Meta, of course, means Facebook, Instagram, WhatsApp, and Threads, among other things.

And just how big are data centers in terms of energy use? More or less France-big while heading toward Japan-big.

And projections into the future? Don’t even go there. There is an increasing proportion of the data center energy that’s sourced from renewables, but that’s not a lot of comfort when it comes along with increased fossil fuels like coal and “natural gas” (methane).

What we need is virtually no fossil fuels.

We’re trying desperately to run up a down escalator.