Epic has confirmed AI tools are being used to develop Unreal Engine code and features, while Tim Sweeney says the Unreal Engine and MetaHuman EULA will change around AI training on outputs.
Image: IGDB
Epic confirms AI is now inside the Unreal Engine development loop
Epic Games has confirmed that AI tools are being used to develop Unreal Engine, turning what had been a public debate over store disclosures and AI policy into a direct question about the future of one of the game industry's most important production pipelines. CEO Tim Sweeney wrote on X that, “Yes, we are now using AI tools to develop code for Unreal Engine,” and Epic separately confirmed to This Week in Videogames that AI tools are being used to develop code and features for Unreal Engine 6 and Unreal Editor for Fortnite.
That confirmation creates immediate tension for developers who build on Unreal. This is not a distant research project or a marketing slogan attached to a prototype tool. According to Epic’s Executive Vice President of Development Marcus Wassmer, AI tools are already “helping across a broad spectrum of development,” with the company using systems such as Claude Code and OpenAI’s Codex in its engineering work. GamingBolt, citing Epic’s statement to This Week in Videogames, reported that Epic said AI is being used to “develop code for Unreal Engine.”
The shift is also notable because Sweeney had previously pushed back on the idea that Epic was using AI tools to improve, update, and build Unreal Engine features. This Week in Videogames reported that in November 2025, Sweeney denied assumptions that Epic was using AI in that way. After Epic’s more recent “huge pivot” toward “maximum AI acceleration,” that answer has changed. The confirmed fact is simple: Epic is now using AI-assisted development for Unreal Engine code. The open question is how deep that practice becomes as Unreal Engine 6 moves from roadmap to daily reality for studios.
The EULA change is about outputs, not a blank check on every game
The legal side of the story is separate from the coding side, but developers are watching both because they meet at the same pressure point: control over what gets fed into AI systems. Game Developer reported that Sweeney said Epic’s EULA currently restricts AI training on outputs from Unreal Engine and MetaHuman, and that Epic is working to remove that restriction from the EULA terms. TechPowerUp reported the same core statement from Sweeney: Epic currently restricts AI training on Unreal Engine and MetaHuman outputs, but is “working to remove this restriction from EULA terms.”
The important distinction is that Sweeney also said individual games remain subject to the terms chosen by their developers, which can allow or restrict AI training. In practical terms, Epic’s stated plan is not the same as announcing that every Unreal project becomes available for AI training. The company is saying the engine-level EULA will no longer impose the current restriction on training from Unreal Engine and MetaHuman outputs, while developers can still set terms for their own released games.
There are still missing details that matter. The sources do not include final EULA wording, an implementation date, enforcement terms, or a plain-language explanation of how Epic will separate general tool outputs from game-specific assets and experiences. For studios, that is the decisive part of the ruleset. A character rendered through MetaHuman, a cinematic exported from Unreal, and a shipped commercial game all sit in different business and legal contexts, even if they pass through Epic’s tools. Until Epic publishes the revised terms, the safest reading is that a change is coming, not that the practical boundaries have been fully defined.
Epic is drawing a line between AI training and Fortnite IP enforcement
The EULA discussion surfaced after criticism of major technology and game companies embracing AI while still enforcing their own intellectual property rights. Game Developer reported that Sweeney responded to a Kotaku piece questioning possible hypocrisy from companies such as Microsoft and Epic, particularly around AI-coded recreations and unauthorized uses of protected games or software.
Sweeney’s response, as described by Game Developer, separates those issues into different categories. He stated that Epic does not let people use Fortnite intellectual property without permission or compensation. At the same time, he said Epic is working to remove the current EULA restriction on AI training from outputs of Unreal Engine and MetaHuman, with developers retaining the ability to set terms for their own games.
That distinction will not end the debate, but it does define Epic’s position. The company is treating unauthorized Fortnite copies as copyright and licensing issues, while treating AI training on tool outputs as a ruleset it wants to loosen at the engine level. Developers should pay attention to that split because it indicates where Epic sees the platform boundary: Epic wants broad AI acceleration around its development stack, but it is not signaling tolerance for unlicensed use of Fortnite itself.
UE6 is being shaped around Unreal, UEFN, and AI-assisted engineering
The Unreal Engine 6 pipeline is the strategic center of the story. GamingBolt reported that Epic offered a look at the next version of Unreal Engine in June, when development lead Marcus Wassmer said the engine would unify Unreal Engine 5 and Unreal Editor for Fortnite. That plan matters because UEFN is tied to Fortnite’s creator economy and live-service scale, while Unreal Engine remains a production backbone for external studios.
According to This Week in Videogames, Epic’s June “Road to UE6” blog said internal teams, including the engine engineering team, were investigating “what works and what doesn’t for code generation.” At that stage, Epic said most of those use cases were not generating mainline Unreal Engine code, though the blog suggested that could come eventually. The newer confirmation shows that Epic has moved further into AI-assisted development since then.
Wassmer described specific areas where AI tools are being used or evaluated, according to GamingBolt and This Week in Videogames. Those include eliminating “non-deterministic cooking bugs,” accelerating shader compression work, monitoring PC stability in Fortnite, identifying driver and hardware-specific issues, automatic root cause analysis, and automatic fix generation. The technical stakes are high because those are not cosmetic workflow conveniences. Cooking bugs, shader compression, and PC hardware variance are the kind of production problems that consume engineering time and create unpredictable pain for players and developers alike.
For developers, the meta-game is clear: if Epic can use AI to reduce long-tail stability problems and speed up engine modernization, Unreal Engine 6 could arrive with a leaner maintenance profile. If the tools introduce opaque failures or inconsistent code quality, the cost moves downstream to studios that rely on the engine. Epic’s public framing is that the upside is large enough to move quickly. Wassmer told This Week in Videogames that Epic is “more concerned about the risks of moving too slowly” than about skepticism toward AI.
Fortnite is Epic’s test server for the highest-risk pieces
Fortnite is central because it gives Epic a live environment where engine behavior, content workflows, hardware stability, and creator tools are all stressed at scale. GamingBolt reported Wassmer’s statement that Epic has started deploying automatic monitoring of PC stability on Fortnite, already detecting issues with specific driver and hardware combinations and features that previously would have taken months to root cause. He also described automatic root cause analysis and automatic fix generation as having started to measurably improve the long tail of difficult PC issues in Fortnite and Unreal Engine.
That is a meaningful claim, but it is still Epic’s claim. The source material does not provide independent benchmarks, failure rates, before-and-after crash data, or examples of AI-generated fixes that shipped to players. What is confirmed is the company’s stated use case: Fortnite is part of the feedback loop for AI-assisted stability work that Epic says also benefits Unreal Engine.
There is a practical reason developers are watching this closely. Fortnite and UEFN operate as a proving ground for workflows Epic later wants to generalize across its tools. If AI-assisted root cause analysis works inside Fortnite’s enormous PC hardware spread, it could become a valuable advantage for studios fighting platform-specific bugs late in production. If it produces false confidence, the risk is subtler: teams may accept machine-generated fixes without fully understanding interactions across rendering, networking, platform certification, or content cooking. In strategy terms, Epic is trying to convert scale into tooling advantage. The question is whether that advantage remains reliable when exported beyond Epic’s own live-service machine.
The launcher experiment shows Epic is willing to ship AI-coded work publicly
The Unreal Engine confirmation follows another signal from Epic: a new Preview version of the Epic Games Launcher that the company described as “vibe coded” because “programming is hard,” according to GamingBolt. Epic claimed the preview launcher is faster and more responsive than the current one, with a new look. In performance charts cited by GamingBolt, Epic claimed the new launcher can cold start 46 percent faster than Steam and restore from the system tray with a 14 percent performance boost over Valve’s storefront.
Those figures are useful, but they should be treated as Epic’s own benchmarking unless independent testing verifies them. The broader point is that Epic is not limiting AI-assisted work to hidden internal scripts. It is willing to connect AI-coded development to public-facing software and performance claims.
That matters for Unreal Engine AI coding because trust in an engine is built through predictable behavior over years. A launcher can be patched aggressively if an experiment misses the mark. An engine release becomes the foundation for commercial games, middleware, custom forks, build systems, and platform submissions. Epic’s willingness to move fast may deliver speed gains, but Unreal developers will want transparency around testing, code review, regression handling, and whether AI-generated or AI-assisted changes are tracked differently inside Epic’s process. None of the provided sources say Epic has published such a technical governance model.
Developers should watch the terms, the tooling, and the upgrade path
For now, there are three confirmed developments. Epic is using AI tools to develop Unreal Engine code. Epic says AI tools are being used across Unreal Engine 6, UEFN, Fortnite, and related engineering work. Sweeney says the Unreal Engine and MetaHuman EULA will be changed so the current restriction on AI training from tool outputs is removed, while individual developers can still allow or restrict AI training for their own games.
There are also major items still unannounced. Epic has not provided final EULA language in the source material. It has not published a detailed breakdown of which Unreal Engine code paths are AI-generated, AI-assisted, or human-written. It has not provided independent performance or stability data for the Fortnite PC monitoring claims. It has not explained how future Unreal Engine 6 releases will disclose AI involvement to studios, marketplaces, publishers, or platform holders.
The practical guidance is to treat this as a platform policy shift in progress rather than a completed migration. Teams using Unreal should review their own game terms once Epic publishes the revised EULA, especially if they want to restrict AI training on their shipped work. Studios planning around Unreal Engine 6 should follow Epic’s UE6, UEFN, and Fortnite engineering updates closely, because Epic appears to be using Fortnite’s live environment and UEFN’s creator pipeline as proving grounds for AI-assisted engine development. The opportunity is faster iteration on old, expensive problems. The risk is that the industry’s most widely watched engine is entering a new production model before developers have seen the full rulebook.