A Neural Network Learns to Play Mario Kart and Yes, We’re Scared

A Neural Network Learns to Play Mario Kart and Yes, We’re Scared

It’s at moments like this that I almost feel overwhelmed. The terms and theories that are used in this little project are kind of beyond my scope mostly because it’s not the type of thing I’m into and as a result a lot of tends to go in one ear and out the other. The good part is that by dint of being a writer I at least get the gist of it. Plus the idea of a neural network ‘learning’ is a big help as well. What’s being shown here is that an artificial network is capable of learning how to manipulate and predict gameplay using Mario Kart. Are we a little nervous that people are trying to improve the neural networks of AI? Maybe a little.

After all this never seems to go well for humans in the movies and in real life it doesn’t seem like it would do any better. Yet humans seem to feel the need to keep creating things that might one day supplant them no matter that it seems like a very ill-advised thing to do. Creation is definitely the goal of any living creature from the smallest to the greatest. It’s one of the best gifts ever given to those that seek to build and bring something new into the world or improve upon something that’s already here, but there’s almost no real line to be drawn when it comes to the limitations of creation.

Many might say that this isn’t so, but that’s a strong sense of morality speaking, not the fundamental truth of creation. One way or another we are bound to create or improve upon what we choose no matter if it’s a great idea or something vaguely horrifying to consider. The reason that people are so uneasy about AI is that they know deep down that the movies aren’t entirely wrong in their estimation that our own efforts to improve our lives through the creation and granting of free will to machines would be the end of us in some way. I won’t say that the first network to be completely free of human influence will target us without mercy, but the propensity for following what humans say and order to the letter is something that machines are notorious for already.

Being a creation of an imperfect being machines are already flawed in their thinking thanks to their creators. It’s difficult enough harboring abstract thoughts as a human, but for an AI to possess such capabilities is downright scary. A neural network that could think like a human, anticipate like one, and in many ways perform like one, could be a real threat that would not be widely accepted among humankind. But then of course we’re talking about Mario Kart here, so it’s not that scary, yet.

The evolution of AI is fast being realized by folks that think it’s one of the greatest things to ever come around. For the rest of us however it’s becoming an uncertainty we kind of have to live with.

Start a Discussion

Main Heading Goes Here
Sub Heading Goes Here
No, thank you. I do not want.
100% secure your website.