AI In Video Games Actually Isn’t That Bad

As AI becomes an increasingly contentious buzzword these days, it is also seeing more use and growth than ever, especially in video games.

Social media is home to heated doomsday debates about AI’s future while executives host talks about the endless potential of the new technology.

So then, where should the game industry draw the line between worry and excitement for AI?

To understand the impact that AI has on video games, it’s important to know that AI has two different meanings in gaming.

AI in Gaming: What Is It?

Traditional AI, also called “game” or “classical” AI, has its origins in games like Space Invader (1978). It involves a set of rules that determine how the game’s moving parts, besides the player-character, behave. Traditional AI allows these parts to operate autonomously and sensibly.

For example, non-playable characters (NPC) operate on code that tells their AI how they should walk through the environment. In games like Death Stranding (2019), the NPCs’ AI will slowly learn to favor even, traversable terrain. Traditional AI is how NPCs display more complex decision-making and thus feel more human.

The modern breakthrough in AI that most people are familiar with is called generative AI. This is the type of AI that was popularized by ChatGPT in 2022. Generative AI uses machine learning just like traditional AI, but the difference here is that new content is generated. Content could include art, dialogue, or voice acting, and it is based off of provided source materials.

An example of generative AI is the social deduction game Hide N’ Seek by Google. Here, multiple AI chatbots were trained on some (unknown to us) English conversation to learn to mimic human text speak. The lines generated by the AI are generally original and created on the spot during gameplay.

Why Does Traditional vs. Generative AI Matter?

The apprehension surrounding AI comes largely from the creative industries of film, music, literature, art, and of course video games. Within these communities, consumers are not always happy to learn when AI was used in media production.

Sometimes this is exacerbated when there isn’t a clear understanding of which AI type was used and in what way:

When used in entertainment media, traditional AI generally acts as a tool to facilitate the creative process. Its use might be to imitate patterns in line art and save time while animating, like in the Tweet above. Or, it could help train bots to act more human-like, like in modl.ai’s game Forge and Fight! 

In contrast, generative AI requires pre-existing art in order to provide an output. To produce dialogue, you might feed it text conversations or famous books. To produce character models, you might use different artwork you find on Twitter or elsewhere. 

As you might imagine, this aspect of generative AI lowers the barrier of entry to any artistic endeavor just as much as it threatens creatives in these industries. Worries about job security and the diminishing value of professional artistic skills are both on the rise.

In a talk at the 2023 Game Developers Conference, developer and Promethean AI CEO Andrew Maximov discusses presents a slide that reads: “If you would have spent $300,000 and 3 years on something everyone around said was important when it really wasn’t, you’d be pissed too.”

Does AI Have a Future in Games?

In terms of consumer desires, the verdict on AI-made video games has not reached a clear consensus. Many agree with Maximov that AI doesn’t raise the ceiling for game development, but the floor.

When more people can contribute to game creation, there’s room for more conversations and developments to be had overall. Collaboration, which is fundamental to the growth of art and ideas, is open to a much wider audience with AI.

Others argue that they enjoy art for its human expression and connection. This party doesn’t see the same value in an NPC that was written by a computer. Typically NPCs are written with underlying goals in mind, like their narrative role and how much time players should devote to character conversation. When an AI writes their dialogue instead, the objective then becomes to allow the player to have infinite conversations rather than intentional ones.

Consuming media that might have uninspired art or steal from others without permission can conflict with players’ internal morals, as well.

What is the Legality of AI in Games?

Steam, one of the most popular video game distribution companies, has consistently denied games with noticeable AI involvement in the past.

Valve, the company that develops Steam, told Eurogamer in 2023 that “our review process is a reflection of current copyright law and policies, not an added layer of our opinion.” This is because copyright laws don’t protect ownership of AI content. Developers using AI could then cause Valve legal troubles for hosting these games on Steam’s storefront.

But in January of this year, Valve updated their policy to accept games with AI generated aspects as long as it is clearly and thoroughly identified as well as “guard-railed” from creating any illegal content during live gameplay.

“Today’s changes are the result of us improving our understanding of the landscape and risks in this space, as well as talking to game developers using AI, and those building AI tools. This will allow us to be much more open to releasing games using AI technology on Steam.”

Valve, January 9 2024.

When it comes to AI’s supposed theft of existing art, AI seems to be winning the battle. Although this is still an ongoing discussion, one court case ruled this February that generative AI is not theft.

AI in Court Cases

In Silverman v. OpenAI Inc. (and two concurrent cases just like it), many authors (including Game of Throne’s own George R.R. Martin) sued OpenAI for allegedly using their work as source material for machine learning.

They failed on all seven accusations, except for the point that OpenAI’s “conduct may constitute unfair practice,” which violates California’s unfair competition law. Otherwise, they were unable to prove that what OpenAI’s chatbot generated was similar enough to their work to qualify as copyright infringement.

Although this is not uplifting news to creatives in any industry, it does coincide with the way that artists inspire one another now. When we create, we are ultimately working off of all of the media we’ve consumed in the past. But if you add your own twist to produce something new, is that really so different from unique AI prompts?

It’s true that AI requires no cultivated skill and that no machine can manufacture the humanity that is so foundational to art. Still, the speculations of theft are not so cut-and-dry.

How Do Developers Incorporate AI in a Thoughtful Way?

If the concerns arise from shrinking job availability and devaluing artistic skillsets, then there’s still a good chance that we don’t have much to worry about.

Generative AI will undoubtedly increase the number of product-driven, uninspired video games in rotation as people try to jump on market trends. Even so, this will not limit others from making the expressive, creative, and very human games that they have been.

The usage of AI to rip off artwork and cut to the end product will be unavoidable, but further adapting AI as a tool to streamline game production for developers would only benefit the games we already enjoy.

Used in this way, AI can be adopted like a calculator.

First, you learn to solve times tables, long division, exponents, etc. on paper. Then as the math gets more complex, use a calculator to speed up the basics you already understand. When a mistake is made with the calculator, you already understand the math enough to fix your work.

In terms of the game industry, AI is a well-rounded tool to make repetitive or logistical tasks quicker for experienced developers. This can lessen the cognitive workload and free up time for more complicated problems, much like the calculator.

AI Is No More Human Than A Tool

At least for now, AI is still in need of constant human correction and is in no state to replace high-level jobs. Instead, it allows for creators to apply their expertise elsewhere on top of helping inexperienced people try their hand at game development.

Overall, AI is certainly no replacement for the human experience in any form of media. It’s normal to worry about what repercussions it might have. However, a closer look at the applications and reception of AI in games gives more reason to be hopeful.

Gamers and developers alike will not stop pursuing human stories in games, and AI can’t get in the way of that.