Let’s be honest: video game AI (artificial intelligence) is not that great. It’s functional (most of the time), but it lacks the flexibility and common sense of real people. You can see the huge popularity of online multiplayer games. multiplayer online games — most computer characters fail in difficult situations, so we turn to other people to find slack. Consider the first five seconds of this video:

While it’s fun to see a cop nonchalantly chasing another cop in GTA, it also instantly shatters the illusion that the characters are thinking beings. However, this sad reality is starting to change, thanks to recent advances in artificial intelligence.

Profitable partnership

Obviously, video games can benefit a lot from AI technology, but AI can also benefit a lot from video games. The Microsoft research project proposed:

AI research and video games are a win-win combination. On the one hand, artificial intelligence technology can provide a solution to the growing need to add realistic, intelligent behavior to the virtual creatures that inhabit the game world. On the other hand, as the game environment becomes more complex and realistic, they offer a range of excellent testbeds for basic AI research.

However, for most of us, artificial intelligence in video games simply means that we get a more realistic gaming experience. But what does that mean, exactly?

How it works


With the power of artificial intelligence, NPCs (Non-Player Characters) will be smarter and more capable of human learning and planning. The NPC’s ability to learn and adapt is correspondingly very powerful. The technology for truly immersive experiences is still a long way off, but there is reason for hope.

There has been a major revolution in the last five years artificial intelligence technologies, using so-called «deep learning» methods, which allow you to efficiently create much larger neural networks that can reason abstractly about complex data. This kind of processing can be made much more efficient using neuromorphic chips. which implement neural networks directly in the hardware. Potentially, video games could deliver AI tasks to cloud servers using dedicated neuromorphic chips. This will prevent game developers from choosing between smart AI and good graphics, as the AI ​​can run on the cloud side, offloading individual computers. It would also allow the AI ​​algorithms used to share knowledge and knowledge between users, making everyone’s experience smarter.

Currently, most video game AI works with decision trees, simple hand-structured flowcharts of behavior that take into account a few variables and use them to decide a course of action. It’s very simple and computationally cheap, but not very reliable — when you get into a situation that the flowchart’s creator didn’t intend, you often end up with nonsensical behavior. In the future, developers will be able to “train” more complex neural networks to respond to much more data in more subtle and more flexible ways.

Google has already created a network that can look at an image and describe in natural language what is happening in it. It’s not that hard to imagine AI characters that can clearly describe what’s going on and why, as well as form complex plans for the future to achieve goals and discuss those plans with the player. That kind of depth and flexibility will open up whole new game genres, and also likely make a lot of you uncomfortable (for a good fictional take on the subject, see NPC).


Everyone has their own opinion about the future of AI in video games. Some people are genuinely happy to play with smarter NPCs, while others raise ethical concerns about randomly creating and destroying intelligent characters. Some wonder about the relationship that could develop between players and NPCs and how it would be decided by society.

However, it is important to understand what artificial intelligence is. is not : we create intelligence, not consciousness — and there is a difference. From where I sit, concerns about morality won’t be relevant to video games for long. Autonomous warfare, on the other hand, is a different story.

Ethical issues aside, what would it be like to play games alongside advanced AI characters?

Realistic Encounters

Emily Short, a game designer known for psychologically complex interactive fiction, has teamed up with a former programmer Sims 3 Richard Evans to create LittleTextPeople, an experimental game studio. Target? Explore the potential of dynamic, social AI in video games.

First studio game Cotillion contains characters interacting with each other to pursue their own goals and motivations, and the game communicates the resulting confusion through text, generating paragraphs of description and dialogue in real time — and everything can be controlled by the player.

“Our project is not so free that you can enter anything, but it is more flexible [чем деревья диалогов],” Short told Edge Online. «Richard’s sim has the potential to have people exhibit fine-grained eccentricity, so it’s possible to write characters who scream when spoken to, or like to talk about themselves and take more conversational initiative than they should.»

In his GDC 2012 keynote, Short gave several examples:

There is a scene where the characters have an accident in the middle of the night and are left on the side of the road by a train driver who knows the area is haunted. So he got mad and left. Elizabeth just got out of the wreck, but she didn’t see the driver leave. So the model of the world asked her the question: «Where did the driver go?» Lucy knows the answer to that question, so she says, «Well, the driver thinks the area has a bad reputation, so he ran away.» Her specific expression of this is customized. Her character is meant to be a little insecure, so she has her own way of saying it. Then there is the standardized response. The doctor says, “Actually, I bet he left because he was afraid we were angry because he crashed our carriage.” And it’s flagged as potentially intended for humor. And Lucy has the opportunity to decide: «Do I find this funny or not?» She does, so she laughs. This model enables us to have a very detailed characterization and characterization in which people express their eccentric traits through the style and form of their conversation.

With the development of this kind of technology, it is clear that gamers have an exciting future. to look forward to.

Computer games

Future AI video games will seriously scare you Angelina 640x360

It makes sense for artificial intelligence to create a home for itself in video games, but how can he develop them? This is what Michael Cook, a researcher at Goldsmiths College and a PhD student at Imperial College London, set out to find out when he was developing a program called ANGELINA.

ANGELINA — short for «New Game Development Lab I named ANGELINA» — develops video games based on various parameters set by Cook.


Pretty creepy, and you probably can’t go wrong for the job of a human developer.

But it is not always the case. Over time, ANGELINA requires less and less information from Cook. In the future, he aims to teach the program to write its own code and develop its own games on its own.

“Now people are sharing ANGELINA games because ANGELINA is cool,” Cook told Polygon. «One day they will share games because games are great.»

What do you think about artificial intelligence in video games? What games or genres of AI-enabled games are you looking forward to? Let us know in the comments below!

Image credit: Pixabay, calloftreyarch/Youtube

Похожие записи