Home / Technology / The difference between video game AI and real AI

The difference between video game AI and real AI




Welcome to TNW’s Beginner’s Guide to AI. This (currently) five-part feature is designed to give you a very basic understanding of what AI is, what it can do and how it works. The guide contains articles on neural networks (published in order), Computer vision, Natural language processing, Algorithms and Artificial General Intelligence.

One of the most common misconceptions associated with machine learning technology is the idea that video games from the 1

970s and 1980s had “artificial intelligence” built in that was able to interact with a human user.

If you’re curious but in a hurry, the video game “AI” in the traditional sense is not what people are referring to in modern times when they talk about artificial intelligence. The “bots” in an online multiplayer game, the enemies in a first-person shooter, and the CPU-controlled characters in old-school Nintendo games are not examples of artificial intelligence, just clever programming tricks.

Artificial intelligence in the form that we are discussing here at Neural includes machine learning systems such as the central neural networks behind Alexa, Siri and Google Assistant. Adobe uses AI to predict what a fix should look like, Google uses it to find a cheap flight for you, and Twitter uses AI to determine which ads to show for you.

At the risk of further confusing things, video game developers often use AI to create video games. For example, the Unreal Engine uses AI to enable real-time rendering of graphics. AI is typically not used to control anything that interacts with the player as it is usually a poor solution to most of the programming problems developers face.

When players think of “the AI” in a game, they are probably not thinking of a set of image recognition algorithms. You think of the CPU controlled enemies who can detect the player’s actions and react. We’ve seen CPU enemies take cover, request assistance, and respond in kind as players use new tactics. Again, this is usually not achieved with artificial intelligence.

There are only so many things an agent can do in a video game. Therefore, it is usually cheaper and easier to code an agent for specific tasks than to train a neural network to control the agent.

Perhaps in the future it will make sense to create AI-supported agents as the size and functions of the games increase in order to explore video game worlds together with the players. One of the most popular forms of developing robust AI systems is by letting go of models in video game worlds. Stars ship and Super mario bros. are among the most popular game worlds for research on machine learning.

However, the purpose of this research has nothing to do with video game development. Researchers observe AI models in game worlds because they are often physics-based, and this helps AI learn how the real world works.

Although there are a few exceptions, we can usually assume any “AI” reference in the gaming world that relates to the control of the CPU over agents – i.e. the enemy orcs in Shadow of war or the AI ​​companion in Fallout 4 – is not actually artificial intelligence. While the developers of both games likely used AI for a myriad of functions in their creation, the games themselves don’t have built-in AI to control NPCs, agents, monsters, allies, or villains.

Published on August 10, 2020 – 21:16 UTC




Source link