The AI learns. When playing Mario, at first, the AI tries to do nothing. It will see that doing nothing is not how the game is supposed to be played – and therefore it will try to move. Eventually, it will run into a wall and see that simply moving is not enough. After several attempts it will jump over the wall; most likely by accident. It will soon learn that jumping is very important to reach the goal – and thus it will jump more often. At some point, it will jump into an enemy. Step by step the AI will make more and more progress until it eventually reaches the goal.
However, one important question rises here: How is this useful for us? I want to use this blog to look at my topic from a less technical standpoint. I want to answer this kind of rising questions that might be unclear – thus, this blog follows a Q&A approach.
So – how is knowing about how an AI learns important?
We can reflect the behavior of the AI onto ourselves. What did we do when we first started to play Super Mario; a game where no explicit controls were given? We explored what we could do. When we ran into a new situation where the old-fashioned “run-to-the-right-approach” did not work, we tried out new things.
The AI mimics our behavior to some degree. Of course, most Super Mario players do not continuously run into the same wall. However, they do try messing with the controls and they do learn the timing for jumps over time – just like the AI does.
Watching the AI interacting with the game might give us hints on how we learn by interaction.
Back to the basics – will AI be the end of mankind?
Artificial intelligence was originally designed as a step-up in Software Development that aims for automated problem-solving. Is it dangerous? In short: No. But it depends.
Most AI is basically just a sequence of code that just does what it is supposed to do. That can be controlling the behavior of an enemy in a video game or recognizing faces from images. Is there a way that AI could be dangerous, then? Yes, but only if it was designed to be dangerous. If it was designed to fulfill a purpose, it will do so.
We can assume that the concept of consciousness or emotions is not fully understood to this day. Thus, these concepts cannot be programmed that easily; rendering AI into a mere tool. However, the concept can be faked, giving the impressions of a possible threat.
Eventually, we will understand consciousness or emotions and we will be able to program them. This will be the time to re-consider this question in a more serious manner.
Will AI influence us in the way we live?
It already does. A regular smartphone can perform auto-correction while texting, it adjusts the brightness of its screen when its environment gets brighter or darker and – of course – reacts to certain voice commands.
Chatbots and smart assistants are on the rise. AI is involved in most modern applications.
Where is the connection to Interaction Design?
Reinforcement learning is learning by interaction. In this case, it is not a user who directly interacts with something, but an AI. I find it still applicable.
What else could AI be used for?
AI is applicable for many areas. Currently, it is a trend to use the pattern recognition abilities of neural networks to build AIs that can predict course changes for cryptocurrencies.
Why Super Mario World?
On the one hand, Super Mario World is a classic. On the other hand, the project that inspired the project also used the game and it does make a good starting point.
These are the questions that I (and a few people who are close to me) came up with. If I missed something important I would like to ask you to leave a comment about it.
