Epic Games’ stated push toward an “AI first” development approach has renewed a familiar argument in games: whether generative AI is a tool developers can simply choose to use, or an infrastructure decision that can reshape how work is done across an industry.
Aakash Gupta, a former Epic producer who worked on Fortnite, has argued that the public dispute is already lagging behind the practical reality. His central point is not merely that one major company favors AI-assisted workflows. It is that Epic also makes Unreal Engine, a development platform used across a large portion of big-budget game production. If AI capabilities become ordinary, integrated options inside the engine’s editor, Gupta believes adoption can become far more routine than the loud public debate suggests.
“The debate happens on stage while the usage happens at the desk,” Gupta wrote while discussing industry adoption.
That is a forceful interpretation, and it should not be mistaken for proof that developers, artists, programmers, performers or players have stopped objecting. They plainly have reasons to care about how these systems are introduced. But it does identify the part of this story worth watching most closely: not only the visible output of generative AI, but the software layer where developers make assets, write systems, test builds and solve production problems.
Why Epic’s pivot is drawing attention
Epic has described a major pivot toward “maximum AI acceleration.” Gupta says the scale and speed of the change are consistent with the company culture he experienced, where a decision from the top can rapidly alter team assignments, reporting structures and project roadmaps.
He pointed to the fast creation of Fortnite Battle Royale, which he says was built in two months, as an example of the pace leadership can demand once it commits to a direction. He also framed the company’s public memo as the visible stage of a change that, in his view, is often already progressing internally.
That background matters because an AI-first plan can mean several very different things. It could describe employees using systems to summarize information or help with repetitive production tasks. It could mean code-generation assistants embedded in daily engineering work. It could cover asset generation, dialogue, balancing tools, automated testing or other workflows. The supplied information does not establish exactly which tasks Epic will require, permit or automate. “AI first” is a broad strategic label, not a detailed account of what every employee or studio using Unreal Engine will do.
Still, the wording signals that AI is being positioned as a core development priority rather than a small experimental side project. That has made the response more intense, particularly following layoffs at the company earlier this year. A technological acceleration strategy always lands differently when creative and technical workers are already concerned about job stability, credit, control and the long-term value of their skills.
Related coverage includes Former Fortnite Producer Says Epic's AI Shift Reflects How the Company Operates.
Unreal Engine is the key part of Gupta’s argument
Gupta’s most consequential observation concerns Unreal Engine. A game engine is the core software environment used to build a game. It can provide tools for rendering graphics, handling physics, organizing levels, scripting gameplay, managing animation and packaging a project for release. Developers customize and extend those systems, but the engine shapes the workflow in which much of that work occurs.
That does not mean a feature inside an engine automatically becomes mandatory. Studios can set internal rules, revise permissions or decline to use a particular tool. Nor does it mean every title built in Unreal has the same pipeline. But tools that are easy to access, sit inside familiar software and appear to save time can exert pressure even without an official requirement. A producer focused on schedules may see efficiency. A worker may see a changed role, a new approval burden or a concern that their work is being reduced to material for a model.
Gupta says Epic has connected models including Anthropic’s Claude and Google’s Gemini to Unreal Engine, with more integrations underway. That statement describes an important distinction: a developer may not have to leave their ordinary working environment to try an AI feature. The experience can become a built-in option rather than a separate product that requires its own purchase, training and justification.
His conclusion is that this distribution path matters more than public statements by individual studios. A studio can prohibit AI-generated art, and performers can challenge AI voice uses. Platforms can require developers to disclose AI use in games. Yet Gupta argues that those responses do not directly control the underlying tools available in a widely used engine.
That does not make public policy, labor action or platform rules meaningless. They can determine what kinds of content are acceptable, what information players receive and which practices create legal or reputational risk. The more careful conclusion is that rules governing outputs and rules governing production tools are different levers. The tension between them will likely define much of the argument ahead.
What “AI agents” and development use can mean
Gupta cited a Google Cloud survey that reported 87% of game developers use AI agents at work. The figure is notable, but it needs context. An AI agent is a broad term, generally used for software that can carry out a task or sequence of tasks with some degree of autonomy. In game production, AI assistance can range from optimization support to dynamic balancing, both uses named in the survey description cited by Gupta.
Optimization is the process of improving how efficiently a game runs. That might involve identifying performance bottlenecks or helping developers inspect complex systems. Dynamic balancing refers to adjusting gameplay variables so that a game’s difficulty, economy or competitive systems behave as intended. Neither term by itself tells us whether a tool is generating player-facing material, advising a human developer or making changes automatically. Those differences matter.
Usage statistics also do not settle the ethical dispute. “Use” can cover a quick experiment, an internal assistance tool or a central production workflow. It does not tell players whether generated work appears in the finished game, how much human review occurs, whether the material was sourced responsibly, or whether workers had a meaningful say over adoption. A high reported adoption figure indicates that the technology is present in workplaces; it does not answer every question about its consequences.
For the people making games, the practical issue is likely to be specificity. Which tasks are being changed? Who approves outputs? When does an assistant become a replacement? Are workers expected to use it? What data is submitted to third-party systems? And how will studios distinguish between AI used behind the scenes and AI-generated material placed in a shipped product?
Disclosure is not just a checkbox
Epic chief executive Tim Sweeney has previously criticized Steam’s AI disclosure requirements, calling them irresponsible and arguing that they make it harder for developers to succeed. That perspective treats disclosure as an obstacle that can burden creators, particularly in a fast-moving technical environment.
Critics see the same labels differently. A disclosure can provide players with information about how a product was made, while also encouraging studios to think carefully before using a technology whose creative, labor and ownership implications remain contested. Neither view can be reduced to a simple argument for or against innovation. The disagreement is about who should carry uncertainty: the developer trying to ship a game, the worker whose role changes, or the player deciding what to support.
Gupta says AI tools can become an easy checkbox in an editor, with many developers selecting the option without engaging in a broader debate. That is plausible as a warning about friction: when a capability is embedded in a familiar tool, it can feel normal before teams have fully considered its trade-offs. Yet it is too absolute to say the controversy is only “for show.” The conflict is public precisely because it affects real decisions about consent, working conditions, creative attribution and consumer trust.
Former Square Enix executive Jacob Novak offered an even sharper reaction to Epic’s direction, suggesting that non-AI workflows could be left behind. His phrasing is a prediction, not an established outcome. It nevertheless captures the fear behind the current moment: not that AI will exist as an optional helper, but that organizations will treat its adoption as the new baseline for productivity.
The practical stakes for studios and players
For studios, tool integration can create a difficult management challenge. Moving quickly may deliver useful prototypes or reduce repetitive work, but speed does not eliminate the need for standards. Teams need clarity about review, security, ownership and responsibility when a tool produces unreliable or inappropriate material. A workflow is not automatically sound just because it is available inside a popular engine.
For players, the most useful information may be more concrete than a broad label. Does a game contain generative AI material? Was AI limited to internal assistance? Is voice or art involved? What human oversight shaped the final work? Those are not questions answered by a single slogan from either side of the debate.
The industry has repeatedly shown that game-making decisions can shape the games players eventually see, even if the connection is not obvious at launch. Debates over a shooter’s presentation, for example, demonstrate why development choices and framing matter beyond the technical feature list, as explored in this look at why a shooter’s framing matters as much as its weapons.
Epic’s AI shift therefore deserves attention beyond the company itself. If AI tools become a standard element of the engine software used by developers, the question will not simply be whether games use AI. It will be whether studios explain their choices, protect the people whose work is affected, and retain human accountability when automated tools enter the production pipeline.
