Epic Games’ still-unannounced Disney game may have been delayed as its development shifts toward a far more aggressive use of generative AI, with a reported internal direction calling for an “AI first” production plan. The reported instruction is unusually sweeping: no more hand-written code, alongside art generated through full AI workflows.
That does not establish what the finished game will look like, whether the directive will remain unchanged, or even which Disney project is affected. But it is a major signal of intent from a company already putting generative-AI capabilities into Unreal Engine. It also turns an abstract industry argument into a practical question: what happens to a large game project when AI is not simply a supporting tool, but becomes the stated foundation of production?
For players, the immediate practical takeaway is modest. A rumored Disney extraction shooter, previously said to be targeting November, was reportedly moved to March before the newer AI-focused direction emerged. Neither the game’s public title nor a release plan is established here, so that March timing should be treated as reported and subject to change—not a firm date to circle on a calendar.
What is reported—and what remains uncertain
The key details are based on accounts from people familiar with the project: a director reportedly told the team to adopt an AI-first approach, explicitly rejecting hand-written code and requiring art to move through full AI workflows. Reports also indicate the move was not universally welcomed internally, with some developers hoping the request would not be taken literally.
Epic’s public position, delivered through Communications Director Jake Jones, is that the company is embracing “AI-accelerated development,” while stressing that Epic developers remain the creative visionaries and drivers. That language matters because it describes AI as acceleration rather than replacement. The reported internal wording, however, appears broader and more absolute. The tension between those descriptions is the story.
There is no supplied evidence that the game has been cancelled, that AI alone caused any specific schedule change, or that an entirely machine-produced game is imminent. There is also no confirmed breakdown of which tools would be used, how approvals would work, what material would be made by generative systems, or how much conventional editing, integration and quality assurance would remain necessary. Those omissions are not minor; they are the difference between a provocative production mandate and a usable development process.
AI-first is not the same thing as AI-assisted
“Generative AI” generally refers to systems that produce new output from prompts or other inputs. In a game-development context, that output could include written material, images, audio-adjacent concepts, code suggestions or technical documentation. “AI-assisted” development usually implies that people retain a conventional workflow and use the tools for particular tasks. An AI-first directive suggests the starting assumption changes: teams are expected to begin with AI-generated output and then direct, revise or integrate it.
That distinction is crucial for code. Hand-written code is not merely text typed into an editor. It is a collection of decisions about game rules, controls, network behavior, performance, player progression, tools and edge cases. A generative system may produce a draft, a function or an explanation rapidly, but a production game still needs developers to determine whether the result fits the project, interacts safely with other systems and behaves correctly under real conditions.
For art, “full AI workflows” can cover a vast range of possibilities. A team could use generated material during ideation, use it for temporary placeholder work, or build it into final assets. Those are materially different practices. A final game also needs visual consistency, readable gameplay spaces, animation compatibility, technical optimization and approval processes. Even if a tool creates an initial image or model direction, that does not automatically answer how it becomes a finished, functional asset in a shipping game.
In other words, acceleration may be real without making human expertise optional. The interesting issue is whether Epic’s reported plan is built around that reality or around a much stronger claim: that the normal creative and engineering starting points can be dispensed with.
Why a Disney project makes the stakes higher
Disney invested $1.5 billion in Epic in 2024, with the companies framing the deeper collaboration around a games and entertainment universe. Disney-owned characters have already appeared in Fortnite as skins, but that investment signaled ambitions beyond individual cosmetic crossovers.
A Disney-linked game therefore carries a different kind of pressure from a small experimental release. The project would need to meet expectations around recognizable characters and worlds while delivering a coherent game experience. If the rumored project is indeed an extraction shooter in the mold of Arc Raiders, the genre label brings its own design demands: players need legible objectives, reliable systems, a clear risk-and-reward loop, and spaces that work repeatedly rather than just looking striking in a single image.
That is not proof that generative tools cannot contribute. It is an explanation of why the production choice invites scrutiny. A recognizable IP does not eliminate the hard work of building a game; it can make consistency and deliberate direction even more important.
The developer concern is part of a wider industry trend
The reported skepticism within the team tracks with the findings of the latest annual GDC survey. In that survey, 52 percent of respondents said generative AI was bad for the game industry, rising from 30 percent in 2025. At the same time, the survey indicated that generative tools are in use across gaming, particularly for research and programming rather than for asset creation.
Those two findings are not contradictory. Developers can use a technology while still worrying about its business consequences, its effect on jobs, the quality of its output or the changes it creates in who gets credited and who makes key creative decisions. A programmer may find a tool useful for research or a starting point for routine work while objecting to an organizational plan that treats generated output as the default for the entire production.
The asset-creation distinction is especially relevant to the reported Epic directive. The GDC results suggest that broad adoption has not automatically translated into broad comfort with using generative AI for final art. That likely reflects the particular creative and production issues surrounding assets: style cohesion, repeated revisions, technical requirements and the role of artists in defining an experience rather than merely filling a content quota.
What “no more hand-written code” could mean in practice
As reported, the phrase is striking because code is one of the places where generative AI is already commonly discussed. Yet the phrase should not be interpreted as proof that people will stop programming. Someone still has to define the task, assess the generated code, decide whether it belongs in the project, debug failures and maintain the larger system. A codebase is not improved just because more code appears faster.
There is a difference between generating a snippet and owning a system. The former can be useful in limited circumstances; the latter means understanding how systems connect, where responsibility lies when something breaks and why a change was made. For a game team, that ownership also matters when design changes arrive—as they inevitably do—because code must evolve alongside player feedback, level needs and production constraints.
Epic’s AI-accelerated framing leaves room for human review and leadership. The reported internal mandate creates doubt about how much room it leaves. Until Epic provides a fuller account of the workflow, observers should resist assuming either extreme: neither that developers will be removed from creative control nor that the reported language is inconsequential.
Delays can create room, but they also create new risk
A delay is often interpreted as a sign that a project needs more time. That is all that can responsibly be inferred here. If the project is being reorganized around new tools and workflows, additional time could be required for training, pipeline changes, internal standards and the difficult work of determining what the technology can reliably handle.
But changing a project’s production philosophy during development can also introduce risk. Teams must learn how to evaluate output, establish who has final authority and make sure new material fits the rest of the game. If the goal is to scale up generation, the need for review does not disappear; it becomes more consequential. More output can mean more material requiring selection, correction and integration.
For audiences, the eventual test will not be whether a game used AI somewhere in its pipeline. Generative tools are already part of the industry conversation and, in some cases, everyday work. The meaningful questions are whether the game is coherent, whether its systems work, whether its art serves play, and whether the people making it retain the authority and time to solve the problems that tools cannot solve on their own.
Epic’s next move will define the story
Epic has a substantial Disney partnership, a major engine platform and an explicit public commitment to AI-accelerated development. That makes the reported Disney-game shift more than a one-project curiosity. It is an early test of how an influential game company intends to translate generative AI from engine features and internal experimentation into a large entertainment production.
For now, the clearest facts are also the narrowest ones: a Disney-associated Epic project has reportedly moved later, an AI-first mandate was reportedly communicated internally, the approach has reportedly drawn employee concern, and Epic says developers remain the creative drivers of its AI-accelerated work. How those positions coexist—especially the promise of human creative leadership alongside a reported rejection of hand-written code and conventional art workflows—will be the question worth watching.
