On the impossibility of time travel as a tool for understanding humans and AI.

While the possibility of time travel can be explored mathematically (and the maths are actually pretty cool — E.g., “Do Tachyons Exist?” — I tend to prefer an argument based in the human condition. Basically, if meaningful time travel backwards is ever possible in the future, then at some point in the future some person will use it to destroy the world in the past. Since we still exist, meaningful time travel backwards is not possible.

The point of this argument is not actually to show the impossibility of time travel, but rather to illustrate the heterogeneous nature of human beings. This is the technological progress version of Rule 34 (if something exists then there is porn about it; if there is no porn about it, there will be soon): If something can be done technologically, then someone, somewhere is going to try to do it.

The implications of this for AI are fascinating and chilling. In the short and medium term, AI and similar technologies have incredible potential for improving human productivity and well being. But we are already seeing the seeds of the long term – China’s social credit ranking system, for example, demonstrates the potential for AI and Big Data tools to create an Orwellian system of autocratic political, economic, and social control. There will always be a human being out there somewhere who, if given the technological tools, will use those tools to exploit and oppress others.