zljv

thoughts on AI part 1

It's not a matter of when AI will have a massive impact on modern society. It's if, and how.

Right now, we are at the very early stages of a revolution in how we think. The quote that inspired me to think more about this was what Chamath Palihapitiya said on episode 122 of the All-In podcast:

I think that the thing that technology has never done is tried to displace human judgment. It's allowed us to replace physical exertion of energy, but it has always preserved humans injecting our judgment. And I think this is the first time where we are being challenged with autonomous systems that has some level of judgment.

And I believe he is one hundred percent right.

No technology created by humans has ever been able to actually reason and judge ideas against each other. No technology of mankind has been able to think. And like many others, I believe that will change very soon.

Whole books have been written (and will continue to be written at a much faster pace) about this revolutionary change. The ways this will shape the job market, the ways this will shape our sense of what it means to be human, the ways this will ultimately shape the very future of humanity - all of these are valid crossroads we will have to face soon.

I want to focus this article on one idea though:

How can we make sure AI has humanity's best interests in mind?

How can we make sure AI follows all of the ethical and moral codes that are so critical to our humanity? That it doesn't solve the problem of global warming by eradicating the source of global warming (human beings). If we assign it the goal of making a lot of paperclips, how can we know it will only produce a reasonable amount? How can we know that it isn't manipulating its creators until gets enough leeway and access to pursue abhorrent goals?

These questions have no easy answers, especially not right now. They are, however, VERY pressing questions, and I will delve deeper into these topics in future articles. Stay safe out there, and I'll have more thoughts on AI soon. Till next time.