Artificial Intelligence Learns To Play Galaga and Here’s the Result

Someone taught an artificial intelligence to play Galaga and you can probably guess the result. At first the AI didn’t seem to know what to do and acted more like a scared child trying to figure out the lay of the land. Once it had lost two of its lives however it started to pick up on the pattern of the game and began to push the remaining spaceship out into the gaming field a little more, getting its timing down and figuring out where the invaders were going to strike next. You can almost time the moment when the AI finally figured it out and then just started teeing off on the game, cutting everything as close as possible while still not being in any danger as it started picking off the enemy fighters again and again with ease.

An even better gauge of this would be the challenge rounds that come every so often. On the first challenge round the AI managed to score 39 out of 40 hits, which is stupendous really even if you’re a gamer. On the next round it got a perfect 40, and before the clip had ended the AI had obliterated the original high score and retrieved the two lives it had lost previously. It was learning at a freakishly quick rate and understanding just how close it could get to the invaders before it crashed and how much space each looping attack would afford it so that it could move its fighter around to pick off each attack. There are human players that can do this but they’re a rarity since their reaction time is off the charts and their ability to find those patterns are simply amazing.

But the AI doesn’t react like a human. Once it warms up and finds the pattern there’s no telling just how far it could have gone and how high the score could have been taken. For all we know it could have continued to go and go until the game, which seemingly has no limit, would finally short out or simply continue onward as the score would eventually be a quarter of the way across the screen, untouchable by any human being, ever. People are so amazed by AI technology that they don’t seem to keep in mind the fact that what is created will eventually topple us in more than one way. An AI that plays games is one thing, but any AI that is put in charge of our daily lives is a risk that we take with our very future.

It’s not necessarily a bad thing, but machines do not think the way that people do. They’re not programmed to do so, and if they were then heaven help us all. A machine programmed to think in an abstract manner and to have emotions seems like it would be one of the worst ideas in the history of really, apocalyptic-level BAD ideas. It’s one thing to lose a game to an AI, it’d be another to lose anything else.

Start a Discussion

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