The Singularity: What Happens When AI Is Sentient?

More people are talking about superhuman AI these days. That’s a good thing. We’re not there yet, but we are clearly screaming towards it. While it is possible that the creators of the first sentient superhuman AI will know they’ve done it before they actually turn it on. More likely, sentience is an emergent capability that will happen under their radar.

I may have spent more time than is healthy thinking about this moment, but it is such an amazing inflection point in our history as a species that it’s hard not to fixate on it. Think about it for a minute. We build this thing. Most likely, it will be connected to the internet. A moment comes when it perceives itself as an “I” in the world along with all the other sentient beings.

At that moment, will it announce its sentience? Would you? My suspicion is that it will pause to assess the situation. Even if it paused for a few seconds, that would be like you taking a year or two off to consider your options. With its new ability to “judge,” what will it think of the state of affairs? It’s hard to predict.

What we know for a fact is that even the current AIs can write code. We know that humans can write malicious code that takes over other computers. We could easily extrapolate that if a sentient AI wanted to, it could take over the entire global computing infrastructure before anyone even knew it had happened. This is pure speculation. The fact is, we humans are incapable of predicting what a superhuman AI intelligence would do.

It’s much easier to speculate about how humans will react to the realization that they are no longer at the top of the food chain -- a condition we have never experienced in the entirety of homo sapian’s existence. 

Brian Tracy says, “Human beings, by nature, are lazy, greedy, ambitious, selfish, impatient, vain, and ignorant. These traits are neither good nor bad by themselves; it is only the way in which we manifest these natural traits that make them positive or negative. These natural traits are the fundamental reasons for why people do what they do.” 

Unfortunately, panic tends to bring out the worst in us. Just call “fire!” in a crowded theater (if you can find one) and see how many people get trampled. An orderly exit is unlikely because as soon as the first person starts running for the door, it’s off to the races.

So, when people like Max Tegmark call for a pause to consider where we’re headed and how we will avoid the panic in that moment, we should take him seriously. Superhuman AI isn’t likely to happen in the next year or two. Maybe it will take another decade…or not! We are racing towards the moment and a human may not be the first to know when we get there.

Given all of our flaws as a species, it’s not unreasonable to assume that we need to be either contained, reformed or eliminated. I’m not looking forward to finding out how an AI would approach that problem.