I've been seeing a lot of articles running around the interwebs saying that AI developers believe AI could kill us all within a ten years. In fact, I saw an article this morning that claimed some believe it could kill us by the end of this decade. To put things in perspective, the end of the current decade is less than four years away. So let me say it now... I doubt this will come to pass.
Now that I've grabbed your attention with my clickbait, let me share my realistic, though less-than-fully-informed perspective. Yes, it's quite possible that AI will do something highly disruptive within the next ten years, but that's a far cry from an extinction event. The only way I see this as even a remote possibility is through the release of a biological agent that's as contagious as measles and as deadly as ebola, or by releasing a toxin into the entire world's water supply simultaneously. Even then, the likelihood of either event wiping 100% of humans from the earth remain pretty low.
Don't get me wrong. I am skeptical about AI. I believe it can do good, but also feel that we're being crazy reckless in our pursuit of it. I believe AI is generally overhyped; both its productivity and its capacity for bringing catastrophe are being exaggerated. With that said though, I certainly agree that we need to slow things down. At our current speed, I believe we're more likely than not to bring about a severe unintended consequence.
No comments:
Post a Comment