Earlier this week, the Bulletin of Atomic Scientists announced that the Doomsday Clock wouldn’t tick any closer to midnight, but that it wouldn’t tick any further away, either. The clock will remain at “three minutes to midnight,” where the Bulletin set it last year after growing concerned about U.S.-Russian relations, nuclear modernization programs, and climate change.
As I wrote in the tech section today, the Clock provides a rare opportunity to talk about “existential risk”: those threats so vast that they could endanger all of humanity. Existential risk is undergoing a bit of a renaissance right now: Nick Bostrom, the Cambridge philosopher who coined the term, is the subject of skeptical magazine profiles; and millions of tech-made dollars have gone into funding “good A.I.” research.
In fact, there’s a sort of debate right now among Silicon Valley technologists: Does climate change or artificial intelligence post a greater existential risk to humanity?
To more climate-attuned forecasters, this can seem a little silly. “Worrying about sentient A.I. as the ice caps melt is like standing on the tracks as the train rushes in, worrying about being hit by lightning,” once tweeted Bret Victor, a former designer at Apple. (He’d go on to write a guide to what the motivated technologist can do about climate change.) Some of the computing industry’s figureheads—among them Peter Thiel and Elon Musk—disagree, or, at least, find A.I. sufficiently worrisome to invest their wealth in stopping it.
What always strikes me about this is that both sides can imagine their own form of historical irony. Imagine two throwaway lines in a circa-2100 historical review:
“Yet even as the planet’s atmosphere reached the point of no return, some of America’s keenest technical minds poured millions into preventing sentient artificial intelligence, a technological feat now believed to be centuries away.”
Or…
“Despite urgent warnings from some of the most talented engineers on the planet about what was to come, the United States government stayed focused on the danger of climate change.”
For me, it demonstrates the limits of conspicuously meta-historical thinking. History is easy to predict in retrospect; to actually live through it is to see thousands of terrifying possibilities that never come to pass (cf. Vice President Palin). I think vastly more wealth should go to stopping climate change than evil A.I.—but maybe wealth should also go toward handling global pandemics, or reducing extreme poverty, or funding America’s sclerotic democratic institutions.
For me, the thought that that history might one day judge our own era is a happy one. But that’s because, if history is still getting written in 2100, it means there will be people to write it.