Watch Your Step
On a recent weekend in the Santa Monica Mountains of Southern California, I found myself in the not unwelcome predicament of having lost the trail. I had lucked into an underused corner of state wilderness, on the slope of a small mountain covered in sumac and yucca, and, instead of backtracking, decided I should push uphill. In short order, the slope became steep enough (and my breathing hard enough) that I decided I should probably turn back. As I made my way down, though, I realized I was on a steep, brambly, rocky slope, with no cell reception. A turned ankle here, while not fatal, would be a major inconvenience. I decided a mantra would get me through. “Care, not speed,” I told myself, as I walked mindfully downhill. “Care, not speed.” The other thing that had been on my mind that day was artificial intelligence. Since the November 2022 release of ChatGPT by OpenAI, AI has become something of a sensation. Media organizations have used it, to varying degrees; it’s come up at work meetings at Earth Island; and I’ve played around with it here and there myself. But now we are on the cusp of an AI explosion, with more and more uses of the technology becoming apparent, and with more and more companies ready to build the chips, data centers, and other infrastructure needed to scale up. (Sam Altman, the CEO of OpenAI, announced early this month that he is looking to raise billions of dollars for an AI chip venture that would include a “global network of fabrication plants.”) The energy requirements of a fully realized, robust artificial intelligence industry is hard to quantify. In part, that’s because we don’t really know the power demands right now, though we know it is already straining power grids in some places. Experts figure it will be much more than bitcoin mining and crypto currency. That’s to say nothing of the water demands of cooling, or the e-waste these centers will produce. And yet we ought not discount AI entirely from an environmental perspective. After all, machine learning is helping researchers better understand freshwater ecosystems, the culture of killer whales, and more. What’s needed then, is the same approach that got me down that mountain: moving with care in mind, not speed. Alas, this is not the way of Silicon Valley, at least not for now. We can only hope the titans of AI can exercise caution, and, where possible, encourage them to do so. Stay safe out there, and watch your step.
|
