The Future of AI

As we move into consideration of where technology will take us, it becomes more and more of a game of speculation. This week’s discussion of AI and the ideas behind the singularity felt deeply in the realm of sci-fi writers and directors. It’s amazing to think that in this day and age — where to be frank, Siri can barely understand basic queries — we are already starting to worry about the creation of a real-life Skynet.

I’m skeptical about the idea that something which stems from the human mind could possibly surpass any conceptions of intelligence. Almost anything created by humans, especially something on the scale that would kick-off the singularity, tends to be flawed in some way. And, I highly doubt that the program would be smart enough to fix itself…So, even if we somehow got to the cycle of self-iteration characteristic of the singularity, wouldn’t there always be some backdoor in the superintelligence, giving humans the chance to shut it down? I’m not sure that, within the limitations of the physical word, there could ever even be an all-powerful AI.

Perhaps more attainable is the idea of an AI that can mimic a human. Here too, I am hung up on the concept that there is inherently something special about the neurochemical reactions that drive human life. It’s really difficult to believe, though it may be true, that something as complex and powerful as the human mind could be replicated by a (very complex) circuit, provided the model were taught well. Given the limitations of current computing, for example the constraints of the number of bits in memory, it seems especially farfetched. Indeed, most everything in tech that attempts to reproduce real life is a mere approximation — e.g. images are represented by pixels in a limited colorspace.

Then again, maybe it’s possible, just beyond the limitations of our current technologies. This gets back to the idea I discussed in my last blog post, that it is difficult for humans to recognize the potential of what’s in front of them, even just in terms of “smart” consumer items. There really is no telling where we are heading in the future with tech, let alone something as complicated and exciting as AI. Right now, we’re left speculating.

 

One thought on “The Future of AI

  1. Great thoughts. While I share your concerns, I worry about an AI that isn’t all-powerful, but powerful enough to manipulate most humans. We’ve shown over and over again that that isn’t that hard to do, and probably not that hard for an “intelligence” that never sleeps and can track so many people simultaneously.

Comments are closed.