artificial intelligence in “GameMaker: Studio”: a brief tools and technology survey

What follows is a brief report regarding options and tools for adding artificially intelligent behavior to GameMaker: Studio, including recommendations:

“AI” Behaviors in Games

“What passes for A.I. is actually simple scripted behaviors such as walking across a platform, following an object, or executing a boss pattern.” – [1]

So we are not talking about standard AI here (e.g. machine learning and expert systems). We are talking about FAR more easily programmed behaviors.

Folks in the GameMaker Studio community generally use the term “AI” for enemy actions such as following, movement patterns, teamwork (swarming), and pathfinding. All of these can be scripted using GML.

Extending GML (Part One)

One can write extensions to GML in languages such as Java [2], and you can compile GML to C++ [3], which both significantly improve computation speed. This would prove imperative to anyone wanting to try using machine learning or a related mathematically demanding approach.

Extending GML (Part Two)

One can buy a GML extension, “AI Massive Pack” [4], which offers off-the-shelf implementations of much game “AI” functionality. Costs around 15 USD. Development appears active.

Other GML extensions exist in this domain, but they do not appear supported (I looked at a few of them, and checked GitHub).

Big Idea from Emily

One could feasibly bring “real” AI into GameMaker: Studio by writing an extension that “learns” a player’s behavior and adapts. This would definitely require Java or C++ to obtain the necessary computational speed. 


To my intended reader, who paid me ten Simbi [5] credits for this analysis, I recommend the following:

  1. Learn GML. You’ll have to do it if you want to cross your barrier. You’ll have to do it to even really understand what is possible with the technology.
  2. Purchase and use the “AI Massive Pack” [4]. This proves far cheaper (in time expended) than implementing these routines by yourself. Using “canned” libraries such as this is standard engineering practice.
  3. Don’t think much about the “Big Idea from Emily” section above until you have overcome your current hurdles.



Post Author: badassdatascience

Leave a Reply

Your email address will not be published.