Newer technology is always a game-changing thing. In the case of video games, it’s literally so. And what other new piece of tech to talk about than what’s all the rage at the moment—artificial intelligence or AI? Companies have been quick to adopt AI for various purposes. Whereas the majority of corporations from Bloomberg to Disney see it as a tool to cut costs, the game development industry has a very different take on all this.
This take is more creative in nature. It’s being used to develop new story arcs, generate new gameplay ideas, and analyze existing player data at a whole other level. Some companies are even experimenting with AI-based NPCs (non-playable characters) where these seemingly harmless folks can now hold entire conversations and retain context throughout their adventures in-game.
We’re at a very early level here. In this post, we’ll focus on a new aspect that can be revolutionized with AI—virtual economies. Let’s hop right into it, shall we?
What are Virtual Economies?
Believe it or not, virtual economies are one of the most lucrative businesses today. A lot of gamers are a part of at least one such economy, sometimes unknowingly. A virtual economy is an in-game trading system. It ranges from the player-based economy of trading posts in classic MMOs such as World of Warcraft and Elder Scrolls Online to Loot Boxes in CS:GO and Valorant.
A strong virtual economy is at the heart of these games. These don’t necessarily have to involve real money transactions. There’s a lot of in-game currency transactions, trading, and item barter involved here. But yes, the most popular forms of virtual economies are tied to real money.
The best case would be CS:GO skins that you can buy from reputed sellers. These are skins that you buy with real money for a nice change of clothes or a weapon loot box. It’s an active, thriving community of transactions and items.
AI in Virtual Economies
Developers are quickly realizing the potential of leveraging AI-based technologies and deep learning models to optimize these in-game virtual economies. The only current problem is the scale. To really scale it up, a developer will need to make some serious investments in AI.
Today, suffice it to say that justifying this cost in front of stakeholders isn’t going to go very smoothly. But the potential is there, and we all know it.
AI has already shown us how we can build better games. Smart NPCs, dynamic procedural environments, choice-based challenges, matchmaking balances, data analysis, anti-cheat, etc., are all very serious real-world use cases where AI is already doing a lot of work or is poised to be a strong factor in the coming years.
Now, let’s shift our focus toward virtual economies.
What exactly can AI do in this space?
Players trade in-game items ranging from armor and weapons in some games to skins and cosmetics in others. These are facilitated mainly with the help of real-world money or some kind of virtual currency that can only be purchased using real money.
With that in mind, here’s how AI can help.
Dynamic Pricing
AI models can be trained to be good at understanding the virtual economy’s market demand. Many financial models are already available, often proprietarily, to do this for real-world markets. These algorithms will be able to maintain pricing stability just by looking at the market trends, user demand, and trading history of these virtual items.
Trend Prediction
With all that data, such algorithms or models can easily predict the rise and fall in demand for certain skins or cosmetics, which will allow the marketplace to adjust the listings and prices accordingly. For example, we know that a lot of people buy bundles and treasures in Dota around The International, only to sell them at a higher price later.
But when you make all this formulaic, suddenly you can predict a lot more trends than just that.
Preventing Fraud
AI can also help in detecting instances of fraudulent transactions and suspicious activities. This is a huge benefit for any marketplace moderator or game developer. The strength of an AI model is the training or inferencing part. This is why AI can predict weather better than a supercomputer. It relies on past data. The more data it is fed, the more accurate it will be.
So, if a company like Valve trains an AI algorithm on fraudulent transactions and suspicious users, it will quickly learn to spot future offenders, which will be an immense help for all the other players.
Leading CS:GO skin selling platforms can develop their own algorithms this way to make sure that their marketplace remains spotless. Such a case has already been implemented by sites such as skinsmonkey.com
Where Do We Go From Here?
The times are changing fast. The future is full of AI, even in the world of video games. Though there’s still time before we can reach a world akin to Cyberpunk 2077, it’s much easier to accept a worldview where we will have AI-controlled virtual economies that can be easily scaled.
AI-driven price predictions, automated price setting, authenticity verification, highly optimized trade algorithms, and so on. Artificial intelligence models and algorithms can bring about a net positive change in the world of in-game economies that so many gamers rely on.
It won’t be an overstatement to say that games will be much better for player transactions and economies with AI in the picture.
Major Concerns in AI Gaming
-
AI Needs Data: AI requires heaps of data and training. Any company that wishes to make a foray here needs to properly label and deconstruct its existing data and then train a model on it. If it’s done in a shallow or incorrect way, such an algorithm will be able to do nothing good for the players.
-
Constant Learning: At its core, AI models are all about constant learning. If they are not allowed to hook into the economies in a friendly manner, then trends will change before they can pick up on them. This is especially challenging because so far, virtual economy data hasn’t really been intended for public use through APIs. Such APIs will be crucial if we want to use AI for continuous learning. This might require major rewrites of the game’s code before they can even be prepared for something like this.
We Wrap Things Up Here
It’s all exciting, that’s for sure. But only time will tell how practical these AI use cases are, or when we finally achieve something valuable for the gamer. Until then, happy gaming.