Epic’s AI MegaJam allows a game without AI. It still needs a finished build

Epic's November AI MegaJam now has a straightforward answer to the question suggested by its name: you do not have to use AI. The organizer's current FAQ says no, while inviting participants to make a playable game in Unreal Engine or Unreal Editor for Fortnite. The useful distinction is between tools that are permitted and work that still has to be finished.

The jam runs November 12–15, with the theme revealed during a November 11 livestream. Epic's announcement encourages AI-assisted workflows, but the current participation instructions should guide entrants rather than an assumption that more automation earns a better result.

The deadline is for something another person can play

Unreal Engine entrants submit a packaged game; UEFN entrants supply a private version code representing a fixed island snapshot. The organizer does not accept a changeable live or playtest code as an equivalent. A short gameplay video is also part of the submission.

That fixed-build requirement is more than administration. A judge needs to assess the same version that was completed within the allowed period. It also forces a team to decide which experiment is the actual game, rather than leave a collection of promising scenes waiting for someone to connect them.

The theme arrives before the build window opens. Tool setup is permitted beforehand, but the organizer says not to begin building the game before November 12 and to declare relevant material made earlier. Teams can work alone or with up to five people; eligibility restrictions remain in the official rules linked from the event page.

A grant is not the same as a prize

Epic lists $65,000 in cash prizes across categories, alongside separate development-continuation opportunities of up to $50,000 for standout projects. The maximum continuation figure is not a payment every participant receives, and it should not be treated as the budget available to build an entry.

The judging categories include theme, gameplay, fun, visuals, audio and level design. Those criteria point back to the finished experience. A team can generate many assets and still struggle to explain what the player is supposed to do; a smaller project with one legible mechanic may be easier to complete and judge.

This is a different question from choosing a production technology for a long commercial project. Our coverage of GameMaker's planned free GMRT exports concerns the cost and runtime choices around ongoing development. A four-day jam adds a much harsher constraint: the toolchain has to work before the deadline consumes the time intended for the game.

Scope is still a creative decision

External assets require the right license and must be listed. AI permission does not remove that responsibility, nor the need to check whether generated code or content actually fits the project. The sensible preparation is to establish a reliable build path and a small enough concept to test repeatedly.

For a team considering the event, the most productive early agreement may be what to leave out. One good interaction, understandable feedback and a clean ending give a judge something complete to assess. An impressive pile of unfinished possibilities does not become a playable game simply because it was produced quickly.