Video Games have become much more than a hobby or a time pass as AI developers have been using video games for their AI’s to learn and EA has done something similar to its self-learning AI that plays Battlefield 1.
EA’s self-learning AI is currently in the experimental phase and as you can see in the video below, the AI learns and adapts from the scratch and reacts according to its learning capabilities.
However, given it is in the experimental phase, it sometimes gets confused but, this shows the future possibilities and what developers can achieve in the future.
In terms of Video Games, this self-learning AI can prove quite useful to give players something challenging to take on as the AI will learn from its mistakes and player’s behavior and will act accordingly.
This will add another layer of complexity to the game’s depth and will provide a true challenge to players who seek it rather than artificially increasing difficulty level in video games.
This self-learning AI is the product of the effort of the SEED “Search for Extraordinary Experiences Division” which is a division of EA. According to technical director Magnus Nordin, “while the human players outperformed the agents, it wasn’t a complete blowout by any stretch”.
However, EA is not the only one to develop something with machine learning capabilities as Ubiosft recently introduced “Commit Assistant”. Commit Assistant is an AI assistant that uses machine learning to spot bugs.
The goal of this AI assistant is to save developer cost and time while trying to detect bugs. Commit Assistant is composed of ten years worth of game code from Ubisoft’s own library.
Currently, Commit Assistant can only detect 6 out of 10 bugs and will only raise a false alarm for a bug or mistake 30% of the time.
What do you think of EA’s self-learning AI playing Battlefield 1? Let us know in the comments.