The big idea news of the month came from the world of Artificial Intelligence. A quick round-up (you'll need twitter access for these):
AI can now explain why jokes are funny. There are caveats here that the specific examples might be cherry picked, but still it's pretty amazing.
AI can create a picture of anything you ask. Very cool pictures too.
On the darker side. Someone casually tried to use AI to bring his childhood imaginary friend to life and it tried to kill him. This one is wild.
Related, a dense post at Marginal Revolution on if/how/when we should pull the plug on an AI that got out of control.
There are really two concerns with AI. The classic is that it runs amok and damages or destroys humanity before we can stop it. I generally placed that risk in the realm of science fiction but now I am not so sure. One of the difficulties is that if it happens, it could happen so fast we couldn't stop it.
The second is that AI gets so good at making our lives easy and convenient that we end up destroying ourselves. If there is an anthropological concept that seems self-evident to me it is that humans need struggle. As the struggle for survival and continuance has faded and comfort, convenience, and security are easy to come by (comparatively) we have invented more struggles -- mostly moral and aesthetic struggles to generate the conflict that provide the justification we need for our lives. To wit, people from a millennium ago, or even a century ago, would take a description of our lives as some sort of utopia, yet we live in constant self-generated fear and rancor.
Now imagine what it would be like if most of the world never had to do anything. Never had to work. Had important decisions made for them. Fulfilled any fantasy at will virtually. We would be so desperate for purpose that we would respond by creating sources of righteous anger that make protests over pronouns seem rational.
Whatever the risks, we are going to have to figure out how to handle it because there is no such thing as halting progress. No matter how much we fear it and regulate it and urge caution, somebody somewhere will push AI to the limit.
