FileHippo News

The latest software and tech news

Artificial Intelligence researchers are always looking for new, innovative applications to push the limits of how their creations think and react. In Germany, Cognitive... The Mario AI Project

Artificial Intelligence researchers are always looking for new, innovative applications to push the limits of how their creations think and react. In Germany, Cognitive Modelling researchers at the University of Tubingen have created an artificial intelligence in the familiar shape of Nintendo‘s premiere plumber – Mario.

mario

The Mario AI Project, subtitled “Learning Autonomous Behaviour in Mario” have uploaded a video to Youtube showing a self aware Mario not only playing the game on his own, but expressing his emotions as they are affected by his environment and by phrases input from the user.

A clever combination of AI and simple speech generation means Mario will not only audibly respond to prompts, but learn from experience and any information told to him. He is also, like so many of us, driven by his emotions and mood, if he is hungry he will look for coins, if curious he will explore his environment. The video shows how he reacts differently in the same situations after learning new things.

For instance at one point Mario is asked what he knows about Goombas, he replies that he knows nothing of Goombas and then goes off exploring, finds a Goomba and jumps on it. When asked after this life changing experience what he now knows about Goombas, he has a new insight, he tells us that they may die if jumped upon. You can’t argue with that.

While obviously a little tongue in cheek, this project is a fun way to begin understanding cognitive modelling and by using an already familiar character, the AI interaction somehow is much more believable. By having Mario mimic our own intelligence, the German researchers say they can learn more about how our own brains work. Is little AI Mario really alive? Well he’s a lot more alive than that Goomba he just jumped on.

Are we setting a dangerous precedent by having game characters with minds of their own? What’s next? Lara Croft arguing with you about the health and safety implications of entering the tomb? Your Call of Duty soldiers seeking peaceful solutions through diplomacy? Pac-Man going on a diet? It’s a dangerous path, but the Cognitive Modelling researchers at the University of Tubingen seem to know what they are doing.

[Image via Russ Payne]