More game developers openly use generative AI despite criticism

Generative AI used to create new images, text and sound based on prompts and training data, has had a contentious history in the game development community recently. While generative AI has become a commonly used tool for those who create user-generated content (UGC), its use as a tool for game developers has come with criticism. In particular, some users question why AI should be used when a human developer could do the job.

Despite this pushback, game developers and publishers have started openly using AI tools. Major games companies such as Unity, Epic Games, Roblox and Ubisoft have all announced generative AI integrations in their development kits. As generative AI is becoming more commonplace in game development, creators appear to be moving past the early criticism in favor of publicly disclosing their use of AI.

Generative AI in games: The criticism

The controversy bubbled up a few months ago when gamers criticized Squanch Games for its use of AI-created art in its game High On Life. Then-CEO Justin Roiland said the art, created by Midjourney, was used to give the world of the game an otherworldly quality appropriate for its alien planet setting. He said they also used Midjourney to create “weird, funny ideas.”

Last year, a report from Good Luck Have Fun revealed that Ninja Theory, among other developers, used a library provided by Altered AI to create vocal performances. Ninja Theory clarified later that it only used the technology to create “placeholder content only to help [Ninja Theory] understand things like timing and placement in early phases of development.”

At the time, both developers received some pushback. Voice actors criticized Altered AI’s use for work that actors themselves could accomplish. Actor Yuri Lowenthal noted to Game Developer that some actors’ motion-capture performances are allegedly being replicated by AI in games without proper compensation.

As for the use of AI art, most of the criticism came from artists who disapproved of generative AI’s use of potentially copyrighted images as training data. In a recent case involving a comic book called “Zarya of the Dawn,” the United States Copyright Office ruled that images created with Midjourney couldn’t be copyrighted. Only the images the comic’s author, Kristina Kashtanova, created could be copyrighted as the Midjourney images “are not the product of human authorship.”

Late last year, artists noticed that all artwork posted to Artstation was opted into sharing for AI training data by default. In response to calls for reversing this policy, Epic Games CEO Tim Sweeney said on Twitter, “We’re not locking AI out by default, because that would turn Epic into a ‘You can’t make AI’ gatekeeper by default and prohibit uses that would fall under copyright law’s fair use rules.”

AI in game creation is becoming typical

However, only a few months later, the conversation around AI tools in game development is not so cloak-and-dagger. At the recent Games Developers Conference, my colleague Dean Takahashi noted that generative AI was something of a darling at the event. Multiple developers were prototyping and displaying ways in which GenAI could be used in a games development.

One of the companies that did was Korean developer NCSoft. The company recently revealed its new AI-based facial animation and sound technology in a trailer for “Project M.” The project, which is created in Epic’s Unreal Engine 5, uses AI to create dialogue and movements based on text and sound given to it by the user.

A few weeks ago, Unity announced it is working on a suite of generative AI tools its game development engine. The announcement video for this new tech shows a developer inputting a text prompt, but does not show what that prompt would have created. Interested users can sign up on Unity’s website.

Another major publisher that went public with its use of AI is Ubisoft. The Canadian company revealed its new AI writing program called Ghostwriter last month. Developed by Ubisoft La Forge, an R&D branch, writers use Ghostwriter to create NPC “barks,” short phrases in-game characters say when players trigger an event. The AI generates the barks, and writers can select ones that fit the event and NPC in question.

One of the first things Ubisoft says in its announcement of Ghostwriter is that it “isn’t replacing the video game writer, but instead alleviating one of the video game writer’s most laborious tasks.” Ben Swanson, Ghostwriter’s creator, gave a talk at GDC about “Natural Language Generation for Games Writing.” The prompt for the talk calls Ghostwriter, “a system that allows the narrative designer to seed their designs with first draft text that can then be accepted, edited or regenerated by writers at the click of a button, unlocking the ability to scale up narrative systems to accompany larger and more complex worlds.”

Finally, one of the largest and most visible game companies to debut AI-based tools is Roblox. Last month, the company rolled out generative AI coding tools for its game creators, including a material generator that simulates lighting and AI code assist that helps developers code faster using natural language. Considering Roblox has a community of millions of developers, that makes generative AI available to far more creators.

The visibility of these tools show that, while game developers aren’t immune to criticism of their use of generative AI tools, it will not stop them from using said tools. Some of them — such as Ubisoft — take the time to preemptively say that generative AI is a tool and can be used just like any other. Stefano Corazza, head of Roblox Studios, summed it up best when he told GamesBeat, “AI is like when the internet started. It’s going to permeate everything.”